Home

The Unscrambler Program Operation

1. 9 5 2 derivative No equivalent As above 6 3 derivative No equivalent As above i 8 Savitsky amp Golay Smoothing Savitsky amp Golay 2 order polynomial only Divide by wavelength No equivalent May be done manually by Compute 11 Correct for reference No equivalent Performed by instrument gt changes Information just copied from data to model file 13 Full MSC Smoothing Moving average Vision only MSC Full 31 Savitsky amp Golay 17 S Golay Derivation 1 order 2 order polynomial only derivative Vision only 32 Savitsky amp Golay 2 4 S Golay Derivation 2 order As above derivative The abbreviations used in the Unscrambler log files are explained in the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices 78 e The Unscrambler Menu Options The Unscrambler Program Operation File Export Model Vision With this option you can export a result file to Vision format This model can then be imported into the software package controlling your instrument All necessary parameters for export to Vision are not always found in the data and or model files and must be entered by you This is explained in more detail in the Chapter NSAS Transformations The Vision Export dialog is more or less similar to the NSAS Export dialog It contains the three sheets General Instrument Parameters and M
2. j Cancel Help Number of Components Suggested no of Components is 3 Bo Variable Residuals Sample Residuals All Numbers of Components C Total Residuals e Inthe Number of Components field choose the number of components for which you want to display Variable or Sample residuals A radio button allows you to toggle between Variable Residuals and Sample Residuals e Alternatively field All Numbers of Components allows you to display Total Residuals for each model component The abscissa of the plot will have as many values as there are components in the model Residuals Dialog PCA Fitting Sheet From the PCA Fitting sheet available only for MCR results you can produce a line plot showing the Variable Sample or Total residuals which summarize the quality of a PCA fitting on your MCR data Use these plots to compare the quality of your MCR fitting to that of a PCA fitting The Unscrambler Program Operation The Plot Menu e 189 e Inthe Number of Components field choose the number of components for which you want to display Variable or Sample residuals A radio button allows you to toggle between Variable Residuals and Sample Residuals e Alternatively field All Numbers of Components allows you to display Total Residuals for each model component The abscissa of the plot will have as many values as there are components in the model Plot Leverage Leverages are useful for the detection of
3. After using the Previous Horizontal PC button the Influence plot is displayed for a model with three PCs Note that in the present case the Number of PCs is changed for the horizontal and vertical axes together another combination would not make any sense 6 Residual variance influence 4 E eevee 05 195 10310 6012 0 BIND i Ja Zug 60120 P DT E S i 10 0 10 0 20 f oazo EONB ARIES qg RSR E N 14 7 8 0 S 100 40 05455 Leverage 0 05 0 10 015 0 20 0 25 0 30 Tutor c pls2 PC 3 3 See also Previous Vertical PC Next Vertical PC Back to Suggested PC and Next Horizontal PC Note To access the Previous Vertical PC Next Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC buttons activate the Source toolbar View Toolbars View Source Next Horizontal PC al The command View Source Next Horizontal PC and its button Next Horizontal PC is part of the PC navigation tool It allows you to switch to the next PC along the horizontal axis of the active plot Note You may also use the lt arrow right gt key as keyboard shortcut for this command 168 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS The command is available in the Viewer for PCA PCR MCR PLS1 PLS2 and three way PLS results on plots of the following types e 2D Scatter plots where
4. A special ASCII file format has been developed to make it easier for you to convert Unscrambler results into other formats of your own choice The file contains all necessary information for prediction and classification See the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices for a description of the ASCII MOD file format When you choose to export this model you enter the dialog Export ASCII MOD In the Name field type the name of the file containing the ASCII MOD data you want to export If you do not remember the name you can click the Browse button to get a list of all available models Dialog Export ASCH MOD PLS1 model Export ASCII MOD Name f Guide T utorial T utorial Result AMO Browse Type Full z Cancel Help Y variable Blue Transformations r Export in The Unscrambler Dout of O selected 9 1 Format Choose between Full or Mini exports from the drop down list in the Type field Mini ASCII MOD exports consist only of the B coefficients from a regression model Full ASCI MOD means that all necessary model results are written to file Results up to and including the PC you specify in the PCs field are exported You can include a list of the transformations that have been performed on the data table from which the results originate To do this press Select in the Transformations field to enter the Select Transformations dialog
5. If your table has an OV layout you may append either Primary or Secondary variables Select the number of variables to append and their type then click OK Samples can be appended as in 2 D data If your table has an O V layout you may append either Primary or Secondary samples Select the number of samples to append and their type then click OK Variables or Category variables can be appended as in 2 D data Edit Append Category Variable Here you may append a category variable which is useful during interpretation of plots and to create groups on which to calculate statistics Note You cannot append a Category variable to a 3 D table with layout OV This option starts the Category Variable Wizard which consists of the following dialogs e Category Variable Wizard Enter Variable Name and Choose Method e Category Variable Wizard Specify Levels e Category Variable Wizard Select Sets Edit Append Mixture Variables This function adds mixture variables to a non designed data table It is not available to designed data tables When choosing this option you are asked for a number of variables to add After the insertion the total number of mixture variables in the table must be three or more Note If you want to append a mixture variable to a designed data table duplicate your table as non design by selecting the menu option File Duplicate As Non Design Then you will be able to append mixture variables to y
6. Kurtosis Flatness X Mean Average value X X Variance Mean square of the deviations from the average Px Sdev Standard deviation spread around the average NEE Trend lines help you to interpret 2D scatter plots You can plot two different trend lines 5 regression line 6 target line View Trend Lines Regression Line Kl or Ctrl E A regression line is drawn between the data points of a 2D scatter plot using the least squares algorithm View Trend Lines Target Line Inserts a target line in your 2D scatter plot The target line is the line with slope 1 0 and offset 0 0 or equation Y X In many cases this line will be the optimal solution eg in predicted vs measured plots View Hotelling T2 Ellipse This function can be used whenever you have a Score plot available in a 2D Scatter plot layout The 95 confidence ellipse is based on Hotelling T statistics Observations found outside the ellipse are potential outliers Note that in a normal situation you would expect about 5 of the samples to lie outside the ellipse Score plot with the Hotelling Ellipse SERA 31 RESULT2 X expl 18 15 Y expl 84 5 a 000 t Qbj358 The Unscrambler Program Operation The View Menu e 159 The Unscrambler User Manual View Uncertainty Test This command is available if you have chosen the Uncertainty Test option when performing a PCA or regression analy
7. The Unscrambler Program Operation The Design Wizard e 107 Design Wizard Define Mixture Variables Design Wizard Define Mixture Variables You are now ready to specify your mixture components Click New to enter a new variable Click Properties to modify an existing variable ID Name LB UB Mix Sum hong I Multilinear Constraints Information Mixture Design Simplex 4 Mixture Variables 0 Process Variables 0 Non Design Variables A Mixture variable is a variable whose quantity is linked to other variables in a mixture Increasing the quantity of A in a mixture of four ingredients will imply decreasing the quantity of B C and or D in order to keep the total mixture amount constant Note You must define at least three mixture variables in this dialog If you are mixing only two ingredients together build a classical non mixture design factorial etc and define the percentage of one of the ingredients in the mixture as a design variable The dialog allows you to add a new or to delete or change an already defined variable by clicking one of the buttons New Delete or Properties Clicking New or Properties leads you to the Add Design Variable or Variable Properties dialog where you can specify the name and properties of the mixture variable In the MixSum field select the total amount of your mixed ingredients between 0 and 100 or between 0 and depending on the mixture component unit
8. Dialog System Setup Users sheet System Setup x Users Import Export Directories Users Guest GU New Delete Switch To I Password required at login gt IT Allow cross user file deletions Supervisor Close Cancel Apply Help The Unscrambler Program Operation System Configuration e 23 Camo Software AS The Unscrambler User Manual e New users are added by pressing New For a single user license you have to delete the existing user before you can create a new one or alternately use Modify on the existing user e An existing user can be deleted by pressing Delete e Select a user from the Users list and press Modify to change the user s name or password e Select a user from the list and press Switch To to log in as that user e Press Supervisor to change the Supervisor s password At the bottom of the sheet the Password required at login tick box allows you to activate the passwords of users other than Supervisor which is always active Tick Allow cross user file deletions if you wish to allow all users to delete files that were created by other users System Setup Dialog Import Export Sheet The supervisor can define how missing values should be handled by default when users import or export data Dialog System Setup Import Export sheet System Setup E x Users Import Export Directories Representation of missing values Unscrambler ASCII files Import
9. In addition to the general information available about the whole plot you may also display specific details regarding one particular point This is done as follows e Rest the cursor close to a data point the point number is displayed e Click on the point a small box containing point number point name and point coordinates is displayed as shown in the figure below 34 e General User Interface The Unscrambler Program Operation Jnscrambler User Manual Camo Software AS Point Details are displayed by clicking on a data point 0 Number 5 Name C1_H1 Abscissa Value 0 817176 3 v C3 H1 Ordinate Value 0 817176 3 2 1 0 1 RESULT1 X expl 27 27 Y expl 91 91 Use of Colors There are three pre set color schemes in The Unscrambler Black background White background and Greytones You can change the color on any of the items of the Viewer This is done through File System Setup Viewer Define colors It is possible to use different color schemes for the screen and the printer For the printer scheme Greytones is recommended if you have a black and white printer You may modify the color scheme and choose your own grey levels Note that also other items than the background and the axis foreground differ in the preset color schemes see below for details The Unscrambler color schemes tem Black background White background Curve 2 Magenta Red je i d
10. Other s field You must also enter the character string used to represent missing data in the ASCII file you want to import in the box String representation of missing data They will be replaced and displayed as m in the Editor 54 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS File Import Lotus You can import data from Lotus 1 2 3 files of WK3 and WK4 type Remember to close the file in Lotus before trying to import it You can enter the range to import manually specifying the upper left cell and the lower right cell eg B2 H34 All cells lying within this rectangle are then imported Do this for the data range and also for the sample and variable names if they exist in the Lotus file Note You must clear the sample and variable names fields if you do not want to import such names File Import Excel You can import data from Microsoft Excel versions 5 7 and 8 Excel 97 workbooks and version 4 0 worksheets The Excel files must have the extension XLS Remember to close the file in Excel before trying to import it When you select to import data from an Excel file in the Import dialog you enter a new dialog Import Worksheet shown below An Excel workbook may contain several worksheets Select the worksheet that contains the matrix you want to import from the drop down menu Sheet name Dialog Import Worksheet Import Workshee
11. Standardization Widely used pre processing that consists in first centering the variables then scaling them to unit variance The purpose of this transformation is to give all variables included in an analysis an equal chance to influence the model regardless of their original variances In The Unscrambler standardization can be performed automatically when computing a model by choosing 1 SDev as variable weights Star Points Distance To Center In Central Composite designs the properties of the design vary according to the distance between the star samples and the center samples This distance is measured in normalized units i e assuming that the low cube level of each variable is 1 and the high cube level 1 Three cases can be considered e The default star distance to center ensures that all design samples are located on the surface of a sphere In other words the star samples are as far away from the center as the cube samples are As a consequence all design samples have exactly the same leverage The design is said to be rotatable The Unscrambler Program Operation Glossary of Terms e 311 e The star distance to center can be tuned down to 1 In that case the star samples will be located at the centers of the faces of the cube This ensures that a Central Composite design can be built even if levels lower than low cube or higher than high cube are impossible However the design is no longer rotatabl
12. This option opens an existing file in a new window If it is a data table an Editor document appears if it is a result file a predefined Viewer appears You enter the dialog Open File as shown below Dialog Open File Look in a The Unscrambler 7 0 El ol Name Type Creator Modified BI NonDesignData 00D Unsc Data GU 19 02 98 10 56 Files of type Data D Cancel I Mine only Information a Print Lab Report Extracted 7 Notes From the drop down list in the field Look in select the folder which contains the file you wish to open The field below will be updated to display all files in the folder of the type specified in the field Files of type You can change the type of file with the drop down list in this field If you know the name of the file you want to open you can type it directly into the field File name otherwise select the appropriate file from the displayed list Press Open At the bottom of the dialog an information field displays information about the selected file You can switch between extracted information i e a brief summary of the file or the notes for the file using the radio buttons below the field If you want a print out of the information press the Print button to the right of the field The Lab Report button directly below it lets you print the lab report for the file File Close This option closes the current data table or result file If you have not previously saved
13. When plotting PCA PCR or PLS results the preview screen on the sheet lets you switch between displaying the plot in one of the 4 corners of the screen and on the entire screen Double click in the preview screen to switch between the options this way you decide whether the plot will appear in a small or large sub view Specific to Loading Weights For Y variables in a PLS model Loadings will be plotted since Loading Weights exist only for X variables Choosing X and Y variables will produce a plot of X Loading Weights and Y Loadings For three way PLS however since there are no loadings Loading Weights are plotted for both X and Y variables Loadings Loading Weights Dialog 2 x 2D Scatter Sheet This sheet lets you specify two 2D Scatter plots which are displayed in sub views 2 and 3 The number of components must be selected for each sub window Specify which components you want to see in the Components field The Unscrambler Program Operation The Plot Menu e 185 Finally make your choice between plotting X Y or X and Y variables in the Variables field Specific to Loading Weights For Y variables Loadings will be plotted since Loading Weights exist only for X variables Choosing X and Y variables will produce a plot of X Loading Weights and Y Loadings Loadings Loading Weights Dialog 2 x 2D Scatter Sheet Three Way PLS Results This sheet lets you specify two 2D Scatter plots which are displa
14. m Parameters Method Principal Component Analysis Y In the Fill Missing dialog you can select the Scope samples and variables of the replacement If necessary click Define to access the Set Editor which allows you to define new sets of samples or variables In the Parameters frame choose the estimation Method e Principal Component Analysis will perform a reconstruction of the missing data based on a PCA model with an optimal number of components of the data selected in the scope field This fill missing procedure is the default selection and the recommended method of choice for spectroscopic data e Row Column Mean Analysis will only make use of the same column and row as each cell with missing data Use this method if the columns or rows in your data come from very different sources that do not carry information about other rows or columns This can be the case for process data You may optionally Scale Data Before Estimating Missing Values by ticking the box at the bottom This is recommended if the variables included in the replacement scope are measured in different units or have different scales Fill Missing Difficult Situations Filling missing values in a satisfactory way requires that the non missing values in the table carry enough information for a reliable estimation The table hereafter lists the most frequent situations and their outcomes Table Various situations and corresponding outcome of Edit
15. 1 0 05 D 05 10 RegrResult X expl 16 47 Y expl 6 12 20 Y variance Residual Validation Variance DI Predicted i Slope Offset Corr 0 055173 7 461658 0 234890 0 230546 9 762251 0 663637 Le 398 S KE 4 3 4 9 6 6 D 5 10 10 4 Measured T T T T T T T T T T PC_00 PC_01 PC_02 FCO 65 70 75 8 0 85 9 0 RegrResult Variable v Tota RegrResutt Y var PC 1 1 For Help press F1 4Wamings RAW GU Plotting from the Viewer Data and results can be plotted in three different ways of which the last two are done from the Viewer e Display a selected part of the data table from the Editor e Display data from any data table or result matrix e Display predefined result plots from an analysis The Unscrambler Program Operation Plotting and Interpreting Results The Viewer e 33 Display Data from a Table or Result Matrix You do this by selecting Results General View An empty Viewer appears giving you access to all data and result files from the Plot menu Display Predefined Plots This option makes use of The Unscrambler s many predefined result plots After each analysis you can choose to see an overview plot of the most important results by pressing View The Plot menu then consists of the appropriate result plots for the type of analysis you have performed You can also access the results plots from File Open at any time Plot Information The Unscrambler
16. Cancel Variables Help fan Variables 52 e Define r Layout OV objects x variables x variables C DV objects x objects x variables m Size Primary variables 4 E Secondary variables fi 3 Samples 17 The Unscrambler Program Operation The File Menu e 81 Camo Software AS The Unscrambler User Manual In the Duplicate as 3D Table dialog specify the scope samples and variables of the data to be converted to 3 D layout In the layout field select either OV or ON Finally define the size of the blocks of variables samples e If the layout is OV type in the number of Primary and Secondary variables e If the layout is O V type in the number of Primary and Secondary samples Click OK when your 3 D table is fully specified Note If the original data table contains variable names or sample names for OV with a structure e g Gum 1 Gum 2 Gum 6 Adh 1 Adh 2 etc the names will be kept in the new 3 D table Otherwise a warning is displayed The Unscrambler xj i The variable names are not built with the required structure for OV2 tables The variable names will be lost File Print Lab Report You can print a form which the laboratory can use when they perform the experiments e g to write down the response values You see common information like sample names who created the design at which time etc in addition to the laboratory report
17. Flatfflat wide ASCII fles m Import gt Export JCAMP DX files fp import Help Cancel System Setup Dialog Directories Sheet The supervisor can move the data directory to a new location by pressing Change on the Directories sheet see the figure below 24 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog System Setup Directories sheet System Setup x Users Import Export Directories Program files C PROGRAM FILES THE UNSCRAMBLER Data files 3 c My Documents The Unscrambler DATA Change Cal Cancel Apply Help Note that the data files are copied to the new location not physically moved This ensures that a backup exists if the location change fails for some reason The previous data directory can be removed manually if desired New User Modify User Modify Supervisor Dialog The dialogs for specifying a new user modifying an existing user and modifying the Supervisor are alike They are accessed by clicking New resp Modify or Supervisor from the Users sheet of the System Setup dialog when you are logged in as Supervisor In this dialog the supervisor specifies the user s name initials and personal password Dialog New User Name OK Val rie Lengard LK Cancel Initials iv Help Password Gg Verify password ee o JV Can Empty Log The name
18. If you choose the second option the values of the source variable will be sorted in increasing order and affected to a range of values between a lower and an upper limit Each interval will then become a new level The Define Ranges for Category Variable Levels dialog allows you to specify those ranges for each of the variables to be converted The Unscrambler Program Operation The Edit Menu e 129 Camo Software AS The Unscrambler User Manual Dialog Define Ranges for Category Variable Levels Define Ranges for Category Variable Levels pA 2 WaterLevel 3 f f With the help of the combo box choose the variable for which you want to specify ranges The number of different values together with the minimum and maximum values of that variable are displayed below the combo box Check these values to help you make your choice in the options that follow Select the desired number of levels between 1 and 20 In the Method field choose whether to divide the actual range of variation from Minimum value to Maximum value into intervals of equal width or specify each range manually Intervals of equal width The system automatically computes the lower and upper limit of each interval defining a new level The levels are named Level 1 Level 2 etc Specify ranges manually If you choose to specify each range manually the list of levels their names and ranges automatically generated can be changed by
19. It is also possible to set the color for a specific item The changes will be shown on the preview screen Dockable Views The Unscrambler shows different kinds of information in dockable views A dockable view is a window that floats on the desktop and which can be glued to the edges of the Unscrambler workspace at wish hence the term dockable Dockable views are toggled on and off in the Window or View menu Dockable Views in the Window menu are Identification and Warning List in the View menu Outlier List The Unscrambler Program Operation Plotting and Interpreting Results The Viewer e 35 Camo Software AS The Unscrambler User Manual Click the title bar of the dockable view to drag it around the screen The shape of the view changes when you get close to the edge of the Unscrambler workplace When you release the mouse button the view is glued to the edge To move it again click inside the docked view and drag it away When you get outside or well inside the edges of the Unscrambler workspace the shape changes again and it has become a floating window Dialogs and Wizards Dialogs and wizards are the elements of the Unscrambler user interface that allow you to give detailed instructions to the program A dialog is a single box that may contain one or several sheets where you choose some options and press OK when you are ready A wizard is a series of dialogs organized in a sequence that lets you proceed thr
20. Model amp subtract Model only Not usec Channel number e Squared channel number ic Squared spectrum if Iw Spectral background information Test samples Omit important variables When EMSC is enabled you must decide which effects to include The options Channel number and Squared channel number model physical effects related to wavelength dependent light scatter variations Chemical effects are included in Squared spectrum For all three options you can choose Not used and the effect will not be included in the transformation If Model only is selected the effect will be included to calculate EMSC parameters By clicking Model amp subtract the effect will not only be included but the effect will also be subtracted from the EMSC corrected spectra It is also possible to include available background spectra in the EMSC calculation by clicking the check box Spectral background information and then clicking on Setup to enter the Background information setup dialog In the Test Samples field you also have to specify the test samples you are going to include in your test set validation in later analyses These samples are not used to find the correction factors but they are MSC or EMSCorrected the same way as the other samples Finally specify the variable ranges that contain chemical information i e not background noise to omit those You may enter several ranges of variables that are not continuous Pr
21. Secondary samples 2 3 removing 14 rows Edit Find Replace Ctrl F This command allows you to find cells containing a given value or sequence of characters and replace the selected value with a new one Edit Find Replace or Ctrl F launches the Replace dialog where you can specify the value to search for launch the search and optionally define a replacement value and perform the replacement The Unscrambler Program Operation The Edit Menu e 127 Dialog Replace Replace Find what f 4anb i S Replace Replace with Ir Replace Al Close Search Bu Columns Match case Help Find entire cells only How to Find a value In the Find what field type in the value or sequence of characters to be searched for By default this field contains the value of the current cell Any combination of digits and characters is allowed e g A51 02 b DSF24 To locate the chosen the value or sequence of characters in a cell hit the Find Next button If the search is successful the cell is marked in the Editor with a black frame or a white frame if the search is occurring in a selected area If no match is found a message is displayed You may change the direction of the search by toggling the Search field either By Columns or By Rows In addition two tick boxes allow for a more specific search e Match case makes a difference between lower case e g a and upper case e g A
22. Set this value lower than 100 res 1 if a constant part of the mixture is not modeled Note You can choose to express the MixSum and the Lower and Upper Bounds either in percentages or fractions This option is called Mixture Component Unit it is found in the menu File System Setup Editor This cannot be done while you are in the design wizard exit to do it then start the design wizard again with File New Design Press the Adjust to Simplex button if you want to work with a simplex shaped experimental region classical Mixture design The settings for lower and upper bounds will be adjusted to create a simplex region Otherwise if your experimental region is not a simplex a D Optimal design will be generated Note Adjust your design to a simplex region if you want to include category variables Check the Multi Linear Constraints checkbox if you want to include multi linear constraints on your variables Define Mixture Variables Next Dialog Once you have added all necessary mixture variables by clicking New or changed their definitions by clicking Properties you may proceed by clicking Next Depending on your choices the next dialog will be the following e If you checked the Multi Linear Constraints checkbox Enter Multi Linear Constraints 108 e The Unscrambler Menu Options The Unscrambler Program Operation e If not Define Process Variables Design Wizard Define Process Variables In this dialog you can define
23. Swap 3 D Layout Modify Toggle 3 D Layouts or Modify Swap Samples amp Variables Modify Transform User defined This function lets you select your own user defined transformation that you can apply on a selection of samples and variables These transformation components can be developed separately and installed on the computer when needed A wide range of modifications can be done by such components including deleting and inserting both variables and samples Dialog User Defined Transformation User Defined Transformation x r Scope Gf Samples F 7 ancel Hel Variables e Selected Variables DI z Define Function Parameter Locate In the Scope field in the User Defined Transformation dialog select the sample and variable sets that define the matrix you want to modify If necessary select Define to enter the Set Editor and define new sets Then select a function of an installed UDT component in the Function drop down menu If a valid function is selected the OK button is enabled and you can click it to complete the transformation Some UDT components may allow additional parameters to be configured possibly depending on the current size of your selected scope Whenever such a configuration is available for the selected function the Parameters button is enabled By selecting it a function specific dialog box will appear allowing you to set various parameters Note however tha
24. To the right of the radio button Name you choose how you want to build up the name character string The first mouse click on a character in the name string indicates the start of a character sequence The next click indicates the end of the name string All characters in between are selected If you want to select characters that are not in sequence hold down the lt Ctrl gt key when you select the characters you want in the label with the mouse Deselect one by using lt Ctrl gt and clicking again Options Dialog General Sheet for Predicted with Deviations and Percentiles Accessed from Edit Options for special plots Predicted with Deviations or Percentiles this sheet lets you define the plot layout Dialog Options General sheet for Predicted with Deviations or Percentiles Options General Cancel Help Plot Header Labels Layout Name Viiv iVivivivivivivivivivivi4 Number Orientation Pasition on File Horizontal C Vertical Options on the General sheet apply to all the data points in the plot The check boxes in the upper right field are used to closer characterize the plots and are enabled and disabled according to which type of plot is activated The available boxes and their characteristics are listed below The Unscrambler Program Operation The Edit Menu e 141 General plot options available for Predicted with Deviations or Percentiles Option D
25. where you will find a list of all available transformations Use the Select button to define which ones to include You can also Select All or Select None For PLS1 models only an extra option allows you to Export in The Unscrambler 9 1 Format It has been introduced to ensure compatibility with some types of external instrument software and uses space delimitation instead of tabulations Note Use option Export in The Unscrambler 9 1 format if you intend to use the PLS1 model for prediction in instrument software from Yokogawa The Unscrambler Program Operation The File Menu e 75 File Export Model U5 Model The Unscrambler saves the model in the old Unscrambler format used by the DOS versions This way other software packages can still read the models you develop in The Unscrambler In the Export U5 Model dialog specify the folder in which the U5 model should be stored in the Folder field Use the Browse button if you do not remember the whole path and have to look at the tree structure In the field Model Name you can enter five characters for the U5 model name The three last letters are used by The Unscrambler to identify different U5 models Check the box Include residuals to include the X residuals Eix and Y residuals Fiy in the exported model file File Export Model NSAS This option exports the result file to NSAS format The model can then be imported into the software package controlling your
26. 31 Select r Special intervals e All variables amp C Every l variables counting from r Interaction and square effects 248 e The Unscrambler Menu Options The Unscrambler Program Operation This dialog operates in the same way as the New Variable Set dialog see p 246 Note Interaction and square effects are disabled in Secondary Variable sets Modify Sample Set Dialog Allows you to modify the properties of an existing Sample Set Lookup the detailed dialog description in New Sample Set Dialog p 245 Modify Variable Set Dialog Allows you to modify the properties of an existing Variable Set Lookup the detailed dialog description in New Variable Set Dialog p 246 Modify Primary Variable Set Dialog Allows you to modify the properties of an existing Primary Variable Set Lookup the detailed dialog description in New Primary Variable Set Dialog p 247 Modify Secondary Variable Set Dialog Allows you to modify the properties of an existing Secondary Variable Set Lookup the detailed dialog description in New Secondary Variable Set Dialog p 248 New Sample Set Modify Sample Set and New Variable Set Modify Variable Set When you press Add in the Set Editor you enter the New Sample Set or New Variable Set dialog The two dialogs are quite similar When you press Properties in the Set Editor you enter the Modify Sample Set or Modify Variable Set dialog These two dialogs are identical to New Sample Set a
27. Dialog Pri X vars Sheet and Sec X vars Sheet These two sheets let you specify the Predictor variables in your tri PLS model Since they are stored in a 3 D array Primary and Secondary X variables are defined separately As an example the Pri X vars sheet is shown below 274 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog Regression Three Way PLS Pri X vars sheet Samples f H Sec X vars Y variables Variable Set All Primary Variables 66 Define Keep Out of Calculation CT Select Weights All 1 0 SDev Iw Include Weights In the Variable Set field choose which variables to use as Primary resp Secondary X variables You can either choose an existing Variable Set from the drop down list or define a new in the Set Editor which you access by pressing Define In the Keep Out of Calculation field you enter any variables which you do not want to include in the analysis indicating ranges by hyphens and separating different ranges by commas for example 1 5 7 10 13 27 You may also select variables to be kept out by clicking Select Weighting is done in the Weights field Only 1 SDev is available as weighting option you select it for all X variables by ticking Include Weights Read more about weighting of 3 D arrays in Chapter Centering and Weighting for Three way Data Regression Three Way PLS Dialog Y variables
28. Dialog System Setup The Editor Sheet System Setup xi General Editor Viewer Missing Data Grid IV Grid lines visible IV Auto adjust columns Include name widths Default numeric format i S Decimals s a C 1 23e 04 Mixture component unit Percentages 0 100 Fractions 0 1 Cancel Apply Help In the field Grid three tick boxes are available e You can turn grid lines on or off by selecting or deselecting the box Grid Lines Visible e The box Auto adjust columns makes The Unscrambler automatically change the width of the columns in the Editor to give enough room for your data Column widths are by default adjusted to fit only the numbers not the variable names e If you tick the box Auto adjust columns you can also choose to adjust the column width to show both numbers and names in full by ticking the last box Include name widths In the field Default numeric format you select the default format of the data values There are three possibilities as shown below Default format of data values Option Pescription Awor fme system decides the format Autofit The system decides the format 12345 6789 Normal layout 1 2345e 04 Scientific notation Use the Decimals spin button to define how many decimal places to use in the latter two options Scientific format allows a minimum of four decimals and normal layout allows a number wi
29. Fill Missing Situation Problem Putcome No missing data No cells to fill Warning is displayed No filling performed Some missing values spread over several No problem Filling is performed samples and variables All values are missing in a row No information available Warning is displayed on that sample Empty row is kept out of calculations More than 50 missing data in the table Risk of unreliable Warning is displayed estimation The Unscrambler Program Operation The Edit Menu e 121 Camo Software AS The Unscrambler User Manual Filling is performed is performed values are missing in the table No information available Warning is displayed at all No filling performed Fill Missing Information Dialog Whenever you open a data table containing some missing values in the Editor an Information dialog appears as shown below Dialog Information upon opening a data table with some missing values Information x The loaded data table contains missing values Not all calculations in The Unscrambler can handle missing data How do you want to proceed Ce Do Nothing C Open fill missing dialogue this time only C Configure automatic filling of missing data T Inthe future do not show this message Choose between the three courses of action e Do Nothing upon clicking OK the data table is displayed in the Editor and nothing more is done e
30. Matrix Plot for each different group in the field Plot Type and in the Sample Groups field choose for which samples you want statistics to be run Design Samples Center Samples Reference Samples or All Samples The first three choices are only available for Statistics run on designed data You can also select sample groups defined using levels of a category variable or subgroups defined by the combinations of levels of several design variables Statistics Dialog Compressed Sheet Here you specify a double plot of the Percentiles displayed as box plot and Mean and Standard Deviation displayed as bars indicating the standard deviation by lines on top of the bars It is designed for simultaneous visualization of several features of location and variability of the data In the Sample Groups field choose for which samples you want the statistics to be plotted Design Samples Center Samples Reference Samples or All Samples The first three choices are only available for Statistics run on designed data You can also select sample groups defined using levels of a category variable or subgroups defined by the combinations of levels of several design variables Statistics Dialog Group Sheet This dialog sheet is only available for Statistics results computed on designed data You specify a quadruple plot showing four Mean and Standard Deviation plots each with a number of groups defined by one of the design variables 206 e The Unscrambl
31. Sample Statistics x EN fee Variable set Cancel 4 Variables 2 Define _ ea Help In the field Scope you select the Variable set for Sample Statistics or Sample set for Variable Statistics on which the statistics will be based For example if you have chosen View Sample Statistics you will access the Sample Statistics dialog where you can select Variable set Sensory descriptors the statistics will be displayed as a table with all your samples as rows and the minimum maximum etc over the Sensory descriptors as columns If necessary use the Define button to access the Set Editor dialog and define a new variable set The results will be displayed in a new Editor which behaves as a slave of the source data Editor For more details see e The View Menu e View Sample Statistics e View Variable Statistics View Sample Names This option shows the design sample names for designed data tables in the Editor The Unscrambler Program Operation The View Menu e 153 View Point Names Applies to designed data only and shows a string of and characters according to the variable levels of the samples in the Editor Variables with several levels are indicated by the level number instead View X Y Variable Numbering Replaces the column numbers in the Editor with an identifier so that you can see which columns belong to the X matrix and Y matrix respectively and which level variable
32. The Status Bar The Status bar at the bottom of the screen displays concise information A short explanation of the current menu option is displayed to the left On the right hand side additional information such as the value of the current cell in the Editor and the size of the data table is displayed Working with Data Tables The Editor The Editor as seen in the figure below handles data in The Unscrambler Each time you open or create a data file its contents will appear in an Editor window You can open several Editors with different contents at the same time switching from one to another as you wish Each time we mention the Editor in this manual and the Help system we mean a window where data or results are displayed in a tabular form The Unscrambler Editor amp The Unscrambler bf_sfc32x 66x31 _ dg Eile Edit View Plot Modify Task Results Window Help Bic dei aa xe A El A a al EE E REES CH In 4 3333 4 2778 Cu IS 4 9722 4 6667 c3 H3 Ia 8 7 0833 4 9722 4 8611 4 7500 macit 3 102 44 100 11 90 78 107 50 93 10 87 32 161 37 151 12 130 80 98 22 93 24 88 06 128 37 115 36 107 63 87 07 82 61 71 24 130 73 112 32 106 26 _1 101 26 92 62 79 83 102 29 96 14 79 83 109 16 97 94 87 04 94 92 83 05 76 95 125 98 103 23 100 61 For Help press F1 Walue 100 31 Size 32x 66x31 R W DG 4 The Editor Basic Notions The Editor consists of a data table made up of rows and columns The intersection of a
33. The Unscrambler User Manual Camo Software AS PCA Dialog Samples Sheet Choose the Sample Set on which you want to base the calculations from the Sample Set drop down list The analysis is performed on the entire chosen sample set If you want to define a new set click Define to launch the Set Editor dialog If you want to exclude some of the samples from the analysis type in their numbers in the field Keep Out of Calculation or click Select to choose them directly from an Editor the Select Samples dialog Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10 13 27 Note How to use Sample Set and Keep Out e In Sample Set you select the general sample population the analysis applies to e InKeep Out you remove one or more of these samples if it turns out that they disturb the results Note Invalid sample range in the Keep Out field will result in an error message as shown below Error message upon invalid sample range The Input alue Is not Accepted Change Value to In the field Frozen Calibration Samples you may type in the list of frozen samples directly or press Select to enter an Editor where you can select the appropriate samples This field is only enabled if you have chosen the validation method Cross Validation The frozen samples are always used in the calibration and are never included in any test sets Thus they can be regarded as especially important
34. This brings up the Print Topics dialog Dialog Print Topics You can print the selected topic or all the topics in the selected heading What would you like to do Print the selected topic Print the selected heading and all subtopics Cancel Choose Print the selected topic if you want to print out only the instructions for the current task If you wish to print out the instructions for the whole tutorial go to the first page of the tutorial e g A Simple Example of Calibration Tutorial AT then click on the Print icon From the Print Topics dialog choose Print the selected heading and all subtopics Illustrations can be printed out one at a time by right clicking anywhere in the secondary Help window where the image is displayed and selecting Print Else you may print out all illustrations for a given tutorial by downloading a PDF file containing the tutorial illustrations from CAMO s web pages www camo com TheUnscrambler Appendices Help Quick Tour Starts a self running demo of The Unscrambler You may choose among several application examples the demo presents the application then shows you how the project is run using The Unscrambler An arrow shows you which menu options are used when running analyses Main results are presented as plots with comments showing how they were interpreted Help Tip of the Day Starts a dialog box where a randomly selected tip is displa
35. This is relevant for Category variables only e Find entire cells only searches for cells which have the requested sequence of digits or characters as exact contents How to Replace a value with another Once you have typed in a Find what value you may proceed with a replacement In the Replace with field type in the new value or sequence of characters Any combination of digits and characters is allowed e g A51 02 b DSF24 However if the requested value is not compatible with the current type of cell e g A51 in a numeric cell an error message will be displayed and no replacement will be made when you hit one of the Replace buttons If the Find what value has already been located with the Find Next button hit the Replace button to replace the value in the current cell If you wish to make the replacement in all cells containing the Find what value hit the Replace All button How to Undo Replace Exit the Replace dialog and use command Modify Undo Replace General features The Close button closes the dialog The Help button launches the Help system on the relevant page 128 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Note If some rows or columns of the data table are selected before using the Edit Find Replace command the search is performed only in the selected area Note only the cells of the data table are affected by this co
36. This is the case for instance for chromatograms e Sensitivity to pure components has a default value of 100 You may tune this value up or down between 10 and 190 by increments of 10 This parameter has an effect on the convergence properties of the algorithm When later viewing your MCR results you will have the possibility to display an MCR Message List recommending changes in Sensitivity whenever necessary to improve the results Note that this option is disabled if you have enabled Initial Guess Task Regression Launches a dialog that allows you to set up the details of your regression model Depending on the nature of the data table displayed in the Editor upon choosing Task Regression the following dialog is accessed 268 e The Unscrambler Menu Options The Unscrambler Program Operation e For a regular 2 D data table the Regression dialog shown hereafter allows you to specify a regression model of type PLS1 PLS2 PCR or MLR e For a3 D data table the Regression Three Way PLS dialog allows you to specify a tri PLS nPLS regression model This dialog is described in section Regression Three Way PLS Dialog Note What follows applies to 2 D data only In the Regression dialog you can set up which data the model will be based on the maximum complexity of the model and how to validate the model Dialog Regression Regression Method PLS1 PLS2 PCR C MLR Samples x variables Y variables Ca
37. Use the command Exact Frame if you need higher precision in your scaling The Unscrambler Program Operation The View Menu e 155 View Scaling Exact Min Max Specify which area of the plot to display You may specify values that are outside the current ranges The Unscrambler does not adjust the scaling to make the plot look nice when this command is used If you want the axes and grid to look nicer try the command Min Max Scaling instead The scaling is done in the Exact Scaling dialog In the field Abscissa Range you enter the minimum and maximum axis value for the horizontal axis Then you enter the minimum and maximum axis value for vertical axis in the field Ordinate Range and click OK View Scaling Exact Frame Select a sub rectangle of the current plot with a frame The selected region is then blown up to fill the entire plot area The Unscrambler does not adjust the scaling to make the plot look nice when this command is used If you want the axes and grid to look nicer try the command Frame Scaling instead View Zoom In Changes the plot scaling upwards in discrete steps allowing you to view a smaller part of the original plot at a larger scale View Zoom Out This option scales the plot down by zooming out on the middle of the plot so that you can see more of the plot but at a smaller scale View Viewpoint You can set a specific viewpoint for 3D plots with the Viewpoint Change command Restore the
38. Use the drop down list which gives you access to the latest used expressions if this is the first time you are using the Compute dialog no formerly used expressions will show in the drop down list 3 Click on the Build Expression button This takes you to the Build Expression dialog wherein you can build your expression using the ready made functions and operators allowed in The Unscrambler The Expression box accepts the expression as a formula of the type X LN ABS X e or S4 S1 S2 S3 or V1 V1 2 SIN V8 V9 where S stands for sample V stands for variable and the number is the sample or variable number in the Editor To build general expressions that are not related to a particular sample or variable use X X stands for the whole matrix defined by the variable and sample set chosen in Scope Note The formula cannot contain mixed references to samples S variables V and X The Unscrambler Program Operation The Modify Menu e 217 The constants operators and functions that are allowed in computations are listed below Table Constants operators and functions allowed in computations pee peton Description Name pescripton Description A oor X2 bn tangent ATAN2 X1 X2 Arcustangent Fess Fro xponential X eAX Multiplication ILN X atural logarithm base e d Division Loo Briggs logarithm base 10 quals to AX X1 X2 aximum value AINT X Integer part of X A
39. Visual Basic Matlab or other UDAs allow the user to program his own analysis methods and use them in The Unscrambler User Defined Transformation UDT DLL routine programmed in C Visual Basic Matlab or other UDTs allow the user to program his own pre processing methods and use them in The Unscrambler Validation Samples See Test Samples Variable Any measured or controlled parameter that has varying values over a given set of samples A variable determines a column in a data table Vertex Sample A vertex is a point where two lines meet to form an angle Vertex samples are used in Simplex centroid axial and D optimal mixture non mixture designs The Unscrambler Program Operation Glossary of Terms e 313 Ways See Modes Weighting A technique to modify the relative influences of the variables on a model This is achieved by giving each variable a new weight ie multiplying the original values by a constant which differs between variables This is also called scaling The most common weighting technique is standardization where the weight is the standard deviation of the variable 314 e Glossary of Terms The Unscrambler Program Operation Index 2 2 x 2D scatter plot 184 2 D 303 2 D data 303 2D scatter plot 175 general view 212 2D Scatter Plot dialog in General View 212 3 3 D 303 3 D data 303 matrix plot 177 3 D layout 307 308 313 3D scatter plot 175 general view 213 3D Scatter Plot
40. and a horizontal rotation of 360 lt and arrows The rotation is executed 5 by 5 vertically and 10 by 10 horizontally If you want a lower rotation or a view from a specific angle you can rotate your response surface 1 by 1 by pressing the Ctrl key at the same time as the arrows Rotate Function used on a Response Surface plot Landscape layout Left hand side view before or after the rotation right hand side view while rotating 36 527 47 510 58 443 69 376 80 309 21 242 Response Surface gt gt eae watts S EE Up Angle Right Angle Pancakes Process PC 3 Y var DLC X var value L C 1 4583 Milk Fat D 1 The Unscrambler Program Operation The View Menu e 157 End the rotate function by clicking on the left mouse button anywhere outside the plot view Note The menu option View Rotate is available for all 3 D plots regardless of the type of data plotted This includes 3 D Scatter plots and matrix plots in the bars and landscape layouts View Projected This command makes a quadruple plot from a 3D scatter plot The upper right window shows the original 3D scatter plot The other windows show the three different 2D scatter projections of the original 3D plot View Plot ID When this option is checked a text line is included at the bottom of the plot with short form plot details Each curve ID is shown in the same color as the corresponding curve in the plot The I
41. groups If you take the option Value of Variable you need to specify which type of variable shall be used for the grouping a Levelled variable an X variable or a Y variable from the drop down list Then you may either type in the variable number in the open text field or click Select to access the Select Variable dialog This dialog is like a small editor showing you the available variables among the chosen variable type and letting you select one of these variables with the mouse In Number of Groups you may tune the number of groups to be generated with the help of the arrows between 1 and 5 groups can be generated The Number of Groups option is not available for Levelled variables in this case the number of groups is automatically defined as one group per level The field Markers Layout contains options available for the layout of the markers in plots The available options are presented in the table below Markers layout options for Sample Grouping Options Description Name Take this option to display the sample names on the plot Use the tick boxes to select the part of the name to be displayed in the plot Number Take this option to display the sample numbers on the plot The basis for the numbering is the current Set Symbol The data points are represented by a symbol e g a small circle which does not give any information about the point Variable Level The level of the sample is used to gener
42. select the X2 plane for which you want to display the regression coefficients In the Components field select the number of components taken into account for the computation of the regression coefficients The regression coefficients are summarized for all components up to and including the chosen number For the loading weights all components up to and including the chosen number are displayed as a line plot in the lower sub view 198 e The Unscrambler Menu Options The Unscrambler Program Operation The dialog displays a preview screen Double click on this to change the size of the plot Plot Regression and Prediction This option produces a double plot of both regression coefficients BW for PCR and PLS B for MLR and predicted values Regression coefficients are used to interpret the relationship between the Y and X variables The plot of the predicted Y versus the measured Y indicates the predictive quality of the regression function For MLR you can include the t values and or p values for the regression coefficient In such cases the plot becomes a quadruple plot In the Regression and Prediction dialog specify whether you wish to plot t or p values in the Plot field Then specify the Y variable for which you want to look at the coefficients and predicted values in the Y variable field and decide whether to include Calibration and or Validation in the analysis Samples field Plot Response Surface This plot visualiz
43. 280 Get Model dialog 280 go to 131 293 Goto dialog 131 Goto dialog 3 D 131 GRAMS files 61 GRAMS32 Import dialog 61 graphical 154 grid lines visible 92 Guided Wave files 63 H hardware requirements 1 help 38 about 301 change license 301 licensee 302 quick tour 300 tip of the day 300 Tutorial exercises 299 Unscrambler help toolbar 297 Unscrambler help topics 297 help button 298 help system 38 access 38 F1 button 38 help toolbar 297 help topics 297 higher order interactions 203 histogram 149 305 plot 176 histogram plot general view 214 Histogram Plot dialog 176 HOIE 203 Hotelling T2 ellipse 159 305 320 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual I amp S 259 260 identification 295 identification window 295 import 50 APC files 61 ASCII files 53 BFF3 64 CLASS PA amp SpectrOn 63 data 41 50 database 56 drag n drop 50 Excel files 55 from spreadsheets 41 GRAMS files 61 Indico files 64 JCAMP DxX files 59 Lotus files 55 Matlab files 61 MVACDF files 62 NSAS files 61 supported formats 50 Tracker files 61 U5 data 53 US results 53 UDI 65 Unscrambler ASCII files 53 Unscrambler data 52 Unscrambler results 51 useful tips 40 User Defined 65 import 3 D supported formats 66 Import 3 D 66 ASCII 68 Excel 68 F3D 69 JCAMP DX 69 Matlab 69 Results 67 supported formats 66 Import ASCII dialog 53 Import dialog 50 51 52 I
44. 3 D data with a 2 way method duplicate it to a 2 D data layout first Additive Noise Noise on a variable is said to be additive when its size is independent of the level of the data value The range of additive noise is the same for small data values as for larger data values Category Variable A category variable is a class variable i e each of its levels is a category or class or type without any possible quantitative equivalent Examples type of catalyst choice among several instruments wheat variety etc Candidate Point In the D optimal design generation a number of candidate points are first calculated These candidate points consist of extreme vertices and centroid points Then a number of candidate points is selected D optimally to create the set of design points Center Sample Sample for which the value of every design variable is set at its mid level halfway between low and high Center samples can be included when all design variables are continuous Closure In MCR the Closure constraint forces the sum of the concentrations of all the mixture components to be equal to a constant value the total concentration across all samples Constraint 1 Context Curve Resolution The Unscrambler Program Operation Glossary of Terms e 303 A constraint is a restriction imposed on the solutions to the multivariate curve resolution problem Many constraints take the form of a linear relationship between
45. Conte Ft ec p valulp valu KSE EAA EAA Plot Type e Line Include Plot F Fales F pValues C Normal Probability I Include Table Y variable h Glossy Significance Testing Method Then specify the Y variable and select a Significance Testing Method from the drop down list for details on the methods see the table in Effects Dialog Overview Sheet 204 e The Unscrambler Menu Options The Unscrambler Program Operation Effects Dialog Effect Details Sheet The plot defined on this sheet shows the Main Effect of a specific design variable or the Interaction between two design variables X variables on one response at a time Choose between a plot of a Main Effect or an Interaction in the Effect field Select the X variable s you want to display the effect of in the X Variables field Dialog Effects Effect Details sheet Effects Overview Response Details Effect Details Cancel Help Effect Main Effect e Interaction Variables Axis Range 1 Dry Matter C 36 0 to 40 0 2 D I wel D bobi Y variable 2 l Shape D Significance Testing Method HoE E Then specify the Y variable to be plotted and select a Significance Testing Method from the drop down list for details on the methods see the table in Effects Dialog Overview Sheet Note In a fractional factorial design the Interaction choice in the Effect field is greyed out if the interactions are confounded On
46. Ctrl M Choose this option to use the mouse to mark objects one by one A click on an already marked object unmarks it Turn the marking mode off by clicking outside the active plot or with the right mouse button within the plot For more general information about marking lookup section Edit Mark Basic Notions Edit Mark Unmark All il or Ctrl U All marked objects in the current plot are de selected The same objects are also unmarked in all other sub views Note If your Viewer contains more than one plot Unmark All applies to objects displayed in the currently active sub view For instance if you are viewing a score plot with marked samples and a loading plot with marked variables choosing Unmark All while the score plot is active will de select the marked samples but will not affect the marked variables Edit Mark Reverse Marking al Choose this option to reverse the current marking status on a plot All unmarked objects variables will become marked and all marked objects variables will become unmarked in one click For more general information about marking lookup section Edit Mark Basic Notions 148 e The Unscrambler Menu Options The Unscrambler Program Operation Edit Copy All Copies all plots displayed in your Viewer in the system of 2 or 4 sub views onto the clipboard You may later on paste the contents of the clipboard into the target of your choice Word Excel and the copied pl
47. Download of The Unscrambler 2 Installation of The Unscrambler 3 Request of an Unscrambler Activation Key 4 Activation of The Unscrambler 5 Definition of user name and password in supervisor login mode A detailed description of each of these steps is given below Downloading The Unscrambler 1 Download The Unscrambler Standalone installer from www camo com and save it into a Temp folder The Unscrambler Program Operation System Requirements e 1 2 Extract the file A folder The Unscrambler Standalone is created Installing The Unscrambler If you are upgrading to a new version it is recommended to uninstall your previous Unscrambler installation prior to installing the new one Un installing your former Unscrambler version will remove the program files only all your Unscrambler data files and model files will be preserved 3 If you have downloaded The Unscrambler from Camo s website double click on file setup exe file found in folder The Unscrambler Standalone to start the installation amp The Unscrambler Standalone 9 5b BAR a File Edit View Favorites Tools Help Czech gt d DO search RE Folders B SZ 2 If mE Address e C TEMP The Unscrambler Standalone 9 5b v Go Folders Name Size Type Date Modified a 0x0409 ini SKB Configuration Settings 25 02 2003 22 34 9 autorun inf 1KB Setup Information 27 10 2005 19 02 Datar cab 24 606 KB Winzip File 27 10 2005 19 02 Slinstmsia exe 1669KB App
48. EE E E SEE TEET 183 PlOt Ladin gS cis 2s cs seicedbes cesessnesceccnnseceseseinshde EEESEEAEEEEEENEEEE EEN NIS dee 185 Loadings Dialog and Loading Weights Dialog A 185 lot E a ET E E EE A TT 186 PlOt E E i 1 E T TE EEEE EREE TEE E 190 Plot MCR O VEVI Wenang a ea E E A E AEA OT EEES 190 Plot Estimated Concentrations 00 0 cc cceeecesccesesceeceeeeeeseeeceeecaeeecneecasescesneseaseeceesseeseeaesaseas 191 Plot Estimated Spectra s ssssevessvecdsieosssavesesndcnatesicess iessen kE eats E EE E iia 191 Plot Regression OVERVIEW sue Seet Ces Zuse EES 191 Plot X Y Relation Outlerg eee cece eessceecesesseeeeseeeecsesseceeseenesaeeeeeecneseesseseeseesaeenes 193 Plot Predicted vs Measured encrier esine nnan EEE E EE 194 Plot Loading Weights sssisciraiiiea nsina ea earnan dca eran e r ETAS 195 Plot Important Variables enenncnennnn e E A NEEN 195 Plot Regression Coefficients isisi peti eerie sarea enr E e EEN E EEEE EEEE 196 Plot Regression and Prediction ccsccssssssscssscsssssssscsseessossscsessssesscecessssesensossssesosasereres 199 PIOE RESPOMSE SULLACE prcn AE EAE RS 199 Plot Response Surface Overview ssssseeseesssrssirrsrsrsrtrsrerstststistrtntstttertste raeentnreseee nterne 202 Plot Analysis of Naam esre erreser EREE aE e R EE EEEIEE 202 Plot E a Se o e E DEES 203 Plot Statisties nnen nnna ar a ir r e E E EEE EEEE S 205 PlOt Classification eege E N E KES erR K 208 Plot Predico EE 209 Plot Line Genera
49. Next and Back again to update the Information field with the latest defined constraints Notes Multi linear constraints automatically lead to a D Optimal design Mixture and Non Mixture One multi linear constraint cannot involve both mixture and process variables No multi linear constraints can be defined in a design including category variables Multi linear constraints are not available in non designed data tables Enter Multi Linear Constraints Next Dialog Once you have added all desired constraints you may proceed by clicking Next You will then access the next dialog Define Non design Variables 110 e The Unscrambler Menu Options The Unscrambler Program Operation Design Wizard Define Model This dialog allows you to include interaction and square effects in a D Optimal design or in a simplex Mixture design Four predefined sets of effects are shown e Mixture Interactions and Squares e Process Interactions e Process Squares e Mixture and Process Interactions In the case of a D Optimal Non mixture design the checkboxes related to mixture variables are disabled Besides interactions and or squares may be disabled depending on the number of defined process variables For more than 6 variables no squares can be included for more than 9 variables no I amp S can be included in the model By clicking the Modify button you enter the Interaction and Square Effects Selection dialog see detailed description b
50. Open fill missing dialogue this time only upon clicking OK the Fill Missing dialog appears and helps you replace the missing values with estimated values e Configure automatic filling of missing data upon clicking OK the Missing Data sheet of System Setup dialog appears and helps you configure automatic replacement of missing values with estimated values in all data tables If you do not want the Information dialog to be displayed in the future tick the box at the bottom of the dial og Note 1 If you choose to Configure automatic filling of missing data the Information dialog will not be displayed in the future Note 2 If the Information dialog does not appear when you open a data table containing missing values you may use File System Setup and configure automatic filling of missing values If you later on turn off that setting the Information dialog will again appear whenever relevant Edit Insert Through this option you gain access to 4 sub options 1 Insert Sample Insert Variable Options and 2 are described in a common section 122 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS 2 Edit Insert Sample or Variable 3 Insert Category Variable 4 Insert Mixture Variables Edit Insert Sample ES See Edit Insert Sample or Variable right below Edit Insert Variable EN See Edit Insert Sample or Variable right belo
51. PCs to be computed 3 Center Data 4 Add Start Noise 5 Issue Warnings Number of PCs In the Num PCs field you define the maximum number of Principal Components to compute Choosing many components will give you a model which explains more of the variations in the data but which also takes more time to build For a first computation we recommend that you choose a rather large number of PCs 1 2 or 1 3 of the total number of X variables for non spectral data around 10 for spectra so as not to miss the optimum in the variance curve Later on you may re calculate your model with fewer components e g one more than the optimal number so that the optimum still shows on the variance curve to limit the size of the model Read more about choosing the number of PCs for tri PLS in Section Maximum number of components from Chapter Three Way Data Analysis Model Size This option is disabled Three way PLS models can only be saved with full size Center Data Ticking the Center Data box makes a model where the data have been centered first i e the average of each variable is subtracted from each of the data values in the variable Centering means that the results can be interpreted in terms of deviations from the average this option is highly recommended You should have a good reason for choosing not to center your data before an analysis Read more about centering of 3 D arrays in Chapter Centering and Weighting for Thr
52. Se Help M Plot Header Labels Layout Name ViV iV iv iv iviv iv iv iv iv iv iv ivi Number Orientation 3 Position on File e Horizontal e Vertical 142 e The Unscrambler Menu Options The Unscrambler Program Operation Options on the General sheet apply to all the data points in the plot The check boxes in the upper right field are used to closer characterize the plots and are enabled and disabled according to which type of plot is activated The available boxes and their characteristics are listed below General plot options available for Mean and SDev Grid Displays a grid allowing you to see the data points coordinates more clearly Plot Header Draws the title of the plot and axis names In the Labels Layout field you choose how markers and labels will appear in the plot Available options are given below Labels layout options Name Select which part of the name to use in the plot Number The basis for the numbering is the current Set Variable Level The level of the sample is used to generate the name for example Indication Name This requires a designed data table Position on file The basis for the numbering is the raw data None No label Orientation field Horizontal or Vertical To the right of the radio button Name you choose how you want to build up the name character string The first mouse click on a character in the n
53. See Weighting 310 e Glossary of Terms The Unscrambler Program Operation Secondary Sample In a 3 D data table with layout O V this is the minor Sample mode Secondary samples are nested within each Primary sample Secondary Variable In a 3 D data table with layout OV this is the minor Variable mode Secondary variables are nested within each Primary variable Segment One of the parameters of Gap Segment derivatives and Moving Average smoothing a segment is an interval over which data values are averaged In smoothing X values are averaged over one segment symmetrically surrounding a data point The raw value on this point is replaced by the average over the segment thus creating a smoothing effect In Gap Segment derivatives designed by Karl Norris X values are averaged separately over one segment on each side of the data point The two segments are separated by a gap The raw value on this point is replaced by the difference of the two averages thus creating an estimate of the derivative on this point Sensitivity to Pure Components In MCR computations Sensitivity to Pure Components is one of the parameters influencing the convergence properties of the algorithm It can be roughly interpreted as how dominating the last estimated primary principal component is the one that generates the weakest structure in the data compared to the first one The higher the sensitivity the more pure components will be extracted
54. Select the combinations you wish to include and press the Add button to include them in the leftmost list under short names You can also Add All Add Int i e Interaction Effects and Add Squ i e Square Effects Use the Remove or Remove All buttons to remove effects from the leftmost list Note If your Variable Set contains more than 62 variables you must push the Select sub set button to define a set with max 62 variables before you can select effects The maximum number of interactions that can be added is 1891 Select Samples Select Variables Dialog This dialog allows you to select samples or variables directly from a data table It is accessed by clicking the Select button in another dialog Typical situations are e From the New Sample Set or Modify Sample Set dialog selecting samples to be included in a Sample Set e From the New Variable Set or Modify Variable Set dialog selecting variables to be included in a Variable Set e From the Samples sheet of any Task dialog e g Statistics PCA etc selecting samples to be kept out from the analysis e From the Variables sheet of any Task dialog e g Statistics PCA etc selecting variables to be kept out from the analysis The dialog consists of a small Editor window that displays the relevant portion of your data table for instance the active sample set and variable set for a PCA and a few buttons The Select Samples dialog is described in detail hereafter The des
55. Split Category Variable Splits one category variable into several new continuous variables There will be as many new variables as there were levels in the category variable For example a category variable with two levels say A and B gets an addition of two continuous variables labeled A and B The samples are coded with 1 and 0 to indicate which level was used originally The figure below illustrates the principle Splitting a category variable This operation is irreversible and only available for non designed category variables Edit Go To bal or Ctrl G Activates a specific cell in the data table You enter the dialog Goto where the field Row lets you choose which sample to go to by pressing the spin button while the field Column works the same way when you want to go to a variable Edit Go To 3 D Activates a specific cell in the data table You enter the dialog Goto where the field Row lets you choose which sample to go to by pressing the spin button while the field Column works the same way when you want to go to a variable If your 3 D table has layout OV the field Column is connected to two additional fields Primary Var and Secondary Var All three fields help you select the desired column in the most flexible way as shown in the figure below The Unscrambler Program Operation The Edit Menu e 131 Dialog Go To for an OV 3 D table If your 3 D table has layout ON the field Row is connected to
56. System Configuration in Practice for help on how to log on as Supervisor It allows you to register i e install a new UDI User Defined Import format or un register a previously registered one with the help of the following commands The Unscrambler Program Operation The File Menu e 69 Camo Software AS The Unscrambler User Manual e File UDI Register New UDI e File UDI Unregister UDI File UDI Register New UDI This command launches the Register New UDI plug in dialog where you can locate the relevant DLL and describe the new UDI for future use Register New UDI plug in dialog Register New UDI K x Static UDI DLL Name CAUDI TestingWCampUDI dl Browse Title Camp Description ft Reaister Cancel Help Pesse __ Gel me In the UDI DLL Name field enter the full path and name of the DLL file containing the UDI you wish to register If necessary click Browse to locate the file Type in a Title for the new UDI this is the name that will identify the user defined Import format for all Unscrambler users Type in a Description make it detailed enough that other Unscrambler users reading this description will understand which Import format you are referring to When this is done click Register a new dialog box informs you that the new UDI has been successfully registered then the Register New UDI dialog is activated again to allow you to register more UDIs When you have
57. TL 22 Supervisor Responsibilities 0 0 eseeseeeeeeesesscneeseceeeseesecnessees caessessesseeeeaesaeeeseaeeeenaees 22 System Configuration in Practice occ cesses cseescesee sescesceeesseesceeesaceeceessaceesaecseeteaeseeeeeaees 23 General User Interface 29 Working E 29 Startups The Main Window eege ege EEN dee E 29 Working with Data Tables The Editor 0 0 cece eceeecseeecseesecceeesseeceecsaeeceees caaeeseessaeeeeeessanaeeaees 30 Plotting and Interpreting Results The Viewer 33 Plotting from the V1 WEN ecien nrerin oia ridode a oi eoa kark 33 Plot tree 34 Pomi RER 34 Use OF COLORS noren ee E EE 35 Dockable E 35 Dalogs and Wizards repite areire r OE TEE E AEA soasensteessdspeeseeusery 36 IR 36 KEE 37 Getting Help in The Unscramble ssis iesarstsss besetavedesieacessenns qtepevasnctd Eo sv EE ea E EEE ESEE rarat 38 The Unscrambler Program Operation Contents e iii The Help System senurnar n n a Eege EE ee EN S 38 Tool EE 38 Communication with Other Software and Instruments 39 Compatibility EE 39 Windows Compabilts 39 Overview of Compatible Format 39 Useful Tips for Import EXpPOtt e isececsesesesecsseapesecpeiecboasesioaseactouesses eysnavebasasteespevaespesvesersaesissinheveeies 40 Exporting Data from The Unscrambler to Other Programs 0 0 cece ceeesnseeeeeeeseeeeeeees 40 Exporting a Model in ASCII MOD Format 41 Exporting a Result Matrix eee eescesccsseecesceseeeesaeeseceeaeesacessaeesaseaeesseeceaeecseeeeaeesn
58. Task section presents the task in a few words section How To Do It gives you detailed instructions e Which commands to use e How to select correct options in the dialogs e How to interpret the results displayed on screen Optimize your Windows The text of the tutorial is displayed in the Main Help window This window by default floats over the Unscrambler workspace You may re size both windows and move them around so as to display the tutorial instructions and the Unscrambler workspace side by side How To Display Illustrations Once in a while you will encounter an enhanced text looking like this Image link Display Image 4003 Clicking on this image link will open a Secondary Help window where the corresponding illustration is displayed This secondary Help window can be re sized and moved around so as to allow you to read the text of the tutorial and look at the illustration simultaneously If you click either on the Unscrambler workspace or the main Help Window the image will be hidden behind If you want the system to remember the size and position of the secondary Help window do not close it it will simply be updated and appear in front again each time you click on a new image link The Unscrambler Program Operation The Help Menu e 299 Camo Software AS The Unscrambler User Manual How To Print Out Instructions and Illustrations From the toolbar at the top of the main Help window click on the Print icon
59. The Unscrambler User Manual Edit Paste 3 D Pasting selections into a 3 D data table is restricted to the unique dimension in the table rows in an ov table columns in an O V table Attempts to paste column oriented selections into an OV table or row oriented selections into an ON table will generate an error message Error message upon illegal Paste The Unscrambler A Can t paste column vectors into a 3 D data table with OV layout Edit Fill Here you can fill the selected range with a value You enter the dialog Fill where you can choose between Specific value Missing data or Level Dialog Fill Fill EN Fill value Specific value Cancel Missing data Help __ Cancel me The Level option is available only if the selected range spans non designed category variables If the value you declare is invalid for some cells these will not be filled Fill is not possible if the selected range contains locked cells for example in a designed data table Edit Fill Missing Here you can fill the missing values in your data table with estimated values that take into account the data structure 120 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog Fill Missing Fill Missing E x Loe Samples pa Cancel W Samples 25 D Define ce Help Variables all Variables 262 D Define
60. The relevant models stored in the current directory are displayed as a list You may click your way to another directory if necessary If you wish to look for models of a specific type e g PCA or PLS1 make your choice in the drop down list of the field Models of Type If you wish to display only models that were created by yourself tick the Mine only box Once a model has been selected in the list more information and new buttons appear in the dialog to help you make your choice The Information field displays details about the contents of the selected model It may be printed out with the Print button The Warnings button leads you to the Text Warning List dialog where you may lookup the wamings generated upon building the selected model The Variance button leads you to the Variance dialog where you may study the variances of the selected model Select the model s you want to use and click Select to return to the main dialog Else click Cancel to return to the main dialog without selecting any model Text Warning List Dialog This dialog is accessed by clicking Warnings from the Get Model dialog The Unscrambler Program Operation The Task Menu e 281 Dialog Text Warning List Text Warning List Waring 1 of 4 Next gt gt Some outlying samples have been detected in Xx T100 If you want more details click Outliers this will display the complete outlier list Else just click Close to go back to the main di
61. Unscrambler Program Operation Troubleshooting under Installation e 21 Un installing your former Unscrambler version will remove the program files only all your Unscrambler data files and model files will be preserved Upgrade from Trial Installation to Permanent License Even if you are not a registered licensee you may download and install a trial installation of The Unscrambler Standalone Once you have tested the program you may find out that you want to purchase it Upon starting The Unscrambler click Details in the Startup Dialog to display the Details dialog showing your license details Dialog Details Details License Details The Unscrambler v9 5b Trial Version 29 Days remaining Click on buy for full version of The Unscrambler For any queries contact support camo com or Cancel support icamo no Contact Camo to place your order Once you are registered as Unscrambler licensee in Camo s customer database you may proceed in upgrading to a permanent license this is done very simply with a new activation key Log in as supervisor see section System Configuration in Practice and click on menu Help Change License to request a new activation key Upon submitting your new request make sure to use your registered licensee e mail ID System Configuration Once The Unscrambler is installed the system supervisor may define new users and specify general options Supervisor Responsibilities The
62. X variables Step and number of Samples are displayed for each file NB Step is the increment in wavelength or wave number between two successive variables The following relationship should be true First X var Step Xvars Last X var Example 64 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Look up file test1 in the dialog snapshot above It has 21 X variables starting at 1600 and ending at 1640 with a step of 2 verifying the relationship 1600 2 21 1640 Once you have selected a file from the list information about the file contents is displayed in the File information field at the bottom helping you check that you have chosen the correct file Multiple selections are NOT possible You may select the additional option at the bottom e Add file name to sample name Click OK to validate your choices and perform the import File Import User Defined Import UDI This option lets you import data from external formats which are not directly supported by The Unscrambler using a DLL Plug in approach This works in a similar way to User defined transformations and analyses see User defined Transformation and requires the desired import format to have been registered in The Unscrambler by your system administrator see Related Topics below The command launches the User Defined Import UDI dialog which lets you choose the desired U
63. Y residuals vs Predicted Useful to detect outliers and or lack of model fit and systematic errors Normal Probability Y residuals A normal probability plot of the Y residuals which is useful to detect lack of model fit Y residuals vs Score A plot of Y residuals vs component scores which is another way to detect lack of fit in the model performed you can use it both for X and Y variables The plot is useful for Influence Plot Squared residuals are plotted against leverage When a regression is outlier detection A plot of the average squared residual for each calibration sample taken over Variance per Sample variables It is useful to detect outliers Variable Residuals A plot of residuals versus sample number for selected variables used to detect outliers and lack of model fit Sample Residuals A plot of residuals versus variable number for selected samples useful for detecting outliers and models with bad fits Sample and Variable Residuals A matrix plot of residuals for samples and variables and another way to detect outliers or lack of fit Choose which samples and variables you want to plot in the Sample and Y Variable fields respectively In addition for PCA PCR and PLS results all the plots require that you select a number of components in the Components field You will then see the residuals for a model consisting of this number of components The Unscrambler Program Opera
64. allows you to enter an SQL query manually 58 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Softwa SQL Query Dialog SQL Query Dialog Data base Path Provider MSDASQL 1 Persist Security Info False D ata List of Tables List of Fields KOSTNADER SATSER DATO LOGBRUKER SATSER DIETT1 LOGGINN SATSER DIETT2 POSTBASE SATSER DIETT3 SATSER NATT SATSER SATSFROKOST REFUSJONSKOSTNADER SATSER SATSFROKOSTPENSJC REISEREGNINGER SATSER SATSFROKOSTBRAKKE REISERUTER SATSER SATSLUNSJ REPRESENTASJON SATSER SATSLUNSJPENSJON REPRESENTERT SATSER SATSLUNSJBRAKKE SATSER SATSMIDDAG STANDARDKJORINGER SATSER SATSMIDDAGPENSJOK UTLANDSKODER SATSER SATSMIDDAGBRAKKE UTLANDSREISER SATSER UTFROKOST UTLANDSSATSER SATSER MAXFROKOST VALUTAER SATSER UTLUNSJ VALUTATYPER SATSER BRAKKESATS CATO OPAC Ink ATC AT Enter Query Run Query Cancel Help From the List of tables select the data table you want to access The List of fields to the right is then updated accordingly If you double click on the table name it will be displayed in the Enter Query area below Type in your query involving available fields in the Enter Query edit box Click the Run Query button to import the data into The Unscrambler To abort the import click Cancel Note If you are not familiar with the Standard Query Language SQL widely used for extracting data from a database you should contact your database administrato
65. and initials defined in this dialog will be traceable in any data or model files that the user creates in The Unscrambler The option Can Empty Log which is active by default may be ticked off made inactive to prevent the user from deleting Log trail information The Log trail is automatically recorded and updated each time a data file is in use The Unscrambler Program Operation System Configuration e 25 Startup Dialog This dialog pops up when you start The Unscrambler either from the Start menu of Windows by double clicking on the Unscrambler icon or by double clicking on the Unscramb exe file Dialog The Unscrambler Startup The Unscrambler Startup OK Cancel Details In this dialog you have to select a user name from the list then click OK If the user you have selected requires a password you enter the Login dialog else the Unscrambler starts right away By default The Unscrambler is delivered with a single user name Guest GU If you wish to add more user names and or apply changes to an existing user name you or your system administrator will have to log in as Supervisor See section System Configuration in Practice for more details on how to do this You may also click the Details button to check the details of your license Dialog Details Details License Details The Unscrambler v9 5b Trial Version 29 Days remaining Click on buy for full version of The Unscrambler For any queri
66. another showing the predicted values In addition you must tick the box Include Table if you want to include a table plot of the predicted and measured values In that case the plot will be displayed in sub view 2 and the table plot in sub view 3 Then select which Y variable you want to see the results from in the Y variable field and the number of components in the Components field Finally select calibration and or validation samples in the Samples field If you have not included the table plot you can change the size of the plot by double clicking in the preview screen in the dialog Plot Loading Weights The loading weights plot is available for PLS1 PLS2 and three way PLS results only It shows which variables are important for the prediction of Y For PLS results X loading weights should usually be plotted together with Y loadings The command Plot Loading Weights launches the Loading Weights dialog see p 185 where you can choose between three different sub view layouts for plotting and select various options type of plot variables PCs Plot Important Variables This double plot combines regression coefficients Bw and loadings loading weights if it is a PLS1 PLS2 or three way PLS model The plot can be useful for detecting the most important variables Plot Important Variables PCR PLS1 and PLS2 Results In the Important Variables dialog select the Y variable for which you want to to display the regres
67. as a whole You must choose between two JCAMP DX formats XYPoints and XYData XYData requires that the distance between each variable s wavelength is the same throughout the whole X Variable Set XYData requires less hard disk space than XY Points Enter information related to the samples in the Samples Info field This information is saved with each sample Export Matlab Files The Unscrambler lets you export data tables to Matlab version 4 or version 5 formats including the sample names row headings in The Unscrambler and the variable names column names in The Unscrambler 72 e The Unscrambler Menu Options The Unscrambler Program Operation Export Matlab dialog Export Matlab EN Sets Sample set jal Samples 19 gt Define Cancel Variable set Help D Variables 20 Di Define Contents File format IV Export sample and variable names The Export Matlab dialog box allows you to specify the data to be exported in terms of the Sample and Variables sets All available sets are listed in the drop down boxes If required new sets can be defined using the Define buttons In the Contents section select the Matlab file version and optionally whether sample and variable names should be exported If this option is selected then these names are stored in separate arrays within the export file as Matlab arrays do not support direct labeling The converted Unscrambler file is saved in a filenam
68. available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices The Unscrambler suggests the optimal Number of PCs based on the residual variance curve It chooses the first local minimum unless later PCs give significantly lower residual variance Note Only models with exactly the same number of variables as in your new data selection will be shown in the Model window Get Model Dialog This dialog is accessed either by clicking Add Model from the Classification dialog or Find from the Prediction dialog It allows you to locate a model with the relevant specifications centered uncentered correct number of X variables 280 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Get Model here in the case of Classification Get Centered Models with 16 X Variables Look jn Test beta2 D EI Creator_ Modified Tutorial G Filmappe 09 03 2003 20 33 af Const PLS2 42D Unse PLS2 05 03 2003 12 22 e EEGEN Unsc PCA 05 03 2003 10 20 Ei Const w cross val 1 Unse PCA 05 03 2003 10 10 Ei Const w testset val Unse PCA 05 03 2003 10 09 Filename Const w AbsToRef CV 11D Models of type fan sl Cancel I Mine only Information PCA A Print Result Name CONST W ABSTOREF CV Warnings irectory C MY DOCUMENTS THE UNSCRAMBLE Creator Dominique Guyot ate 05 03 2003 10 20 Software Version v8 0 Beta Calibration Method PCA e Extracted Notes Variance
69. changes to the document The Unscrambler asks if you want to save the document before you close it Note If you close a window without saving you will lose all changes you did not save previously This includes invisible changes in the file log which are made automatically every time you run an analysis 48 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS File Save II or Ctrl S Saves the active file displayed in an Editor or a Viewer with the name location and file format you previously set in the Save As dialog When you save a document for the first time The Unscrambler displays the Save As dialog If you want to change the name location or file format of an existing document you must choose the File Save As command File Save As Lets you save the contents of the current window as a new file without destroying the original file Specify the new location name and type of file File Save Plot Access this option from the context sensitive menu right mouse click Save Plot or from File Save Plot when a plot is displayed in the Viewer This option leads you to the Save Plot dialog where you can save a plot as an image file in five different formats bmp jpg gif png and tif Example The X and Y loadings plot of a PLS regression analysis can be saved by right clicking on the plot and choosing the desired file format in t
70. command is useful after you have used Add Plot and Scaling View Scaling You can change the plot by scaling its axes to fit the range you want You may scale the plot in 4 ways 1 Select the minimum and maximum values on each axis Min Max Scaling 2 Use a frame to select the desired plot range Frame Scaling 3 Select the exact min and max values on each axis Exact Min Max Scaling 4 Use an exact frame Exact frame Scaling Use Autoscale to display the plot as it was originally View Scaling Min Max You enter the dialog Scaling where you must specify which area of the plot to display You may specify values that are outside the current ranges Dialog Scaling Abscissa Range From 2 77 To 2 09 Cancel Help Ordinate Range E From EE To H 80 The Unscrambler actually adjusts the plot range to make it look nice when you use this option Use the command Exact Min Max if you need higher precision in your scaling In the field Abscissa Range you enter the minimum and maximum axis value for the horizontal axis Then you enter the minimum and maximum axis value for vertical axis in the field Ordinate Range and click OK View Scaling Frame E Sech With this command you can select a sub rectangle of the current plot with a frame The selected region is then blown up to fill the entire plot area The Unscrambler actually adjusts the plot range to make it look nice when you use this option
71. confounded with two variable interactions Resolution IV Main effects are not confounded with each other Main effects are not confounded with two variable interactions Two variable interactions are confounded with each other Resolution V Main effects are not confounded with each other Main effects are not confounded with two variable interactions Two variable interactions are not confounded with each other Two variable interactions are confounded with three variable interactions Three variable interactions are confounded with each other Design Type Next Dialog Once you are satisfied with your choice press Next to enter dialog Design Details Design Wizard Design Details This dialog allows you to specify details related to replicates Center samples and Reference samples 116 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Design Details Choose the number of replicates for the whole design and the number of center samples to be added Design Number of Replicates E Number of Center Samples 2 a Click New to define new reference samples m Reference Points Name Replicates H KE Gesi Jl ge In the field Design you choose the number of replicates and center samples The Number of Replicates indicates the number of times the main experiments are run Increasing the number of replicates gives more precise estimates of the effects due to better coverage of
72. design Then press the Generate button to start generating D optimal design Condition Number to correctly explore the experimental region 321 44 Choosing a solution with a lower condition number will result in reducing the explored region Select a Design to Use 18 Design Points Condition No 357 41 16 Design Points Condition No 379 32 lt Back Next gt Cancel Help In the upper field choose the number of design points that you want by operating the spin button The default number of design points is e Screening designs no square terms P 4 e Optimization designs square terms P 6 where P is the number of design variables interaction terms and square terms in the model This default number allows a correct estimation of the experimental error However if you really need to perform fewer experiments the system will let you decrease this number down to P 1 for a screening design and P 4 for an optimization design On the contrary if you can make do with more experiments increase this default number to gain accuracy Note When choosing the number of design points keep in mind that this number does not include center and reference samples and must be multiplied by the number of replicates to obtain the real number of experiments you will have to produce After selecting the number of design points press Generate to let the system start the D optimal algorithm The system generates four
73. dialog The Unscrambler AN Do you want to save your PCA result Choosing Yes in this dialog has the same effect as selecting File Save As when the results file is opened in the Viewer If you choose No the results are lost and can only be re created by running the same task once again Task Clustering Launches an unsupervised classification of samples using the K means clustering method The samples will be grouped into K user determined number clusters based on a specific distance measurement so that the sum of distances between each sample and its cluster centroid is minimized In the Clustering dialog you can specify the technical details of the computations The Unscrambler Program Operation The Task Menu e 255 Camo Software AS The Unscrambler User Manual Dialog Clustering KMeans Clustering KMeans x Samples Variables Sample Set Keep Out of Calculation Cluster Parameters Number of clusters 4 Number of iterations po Distance type Euclidean b General Options In the Cluster Parameters frame at the bottom of the dialog box you specify the Number of clusters the samples should be divided into and the Number of iterations to be performed before the results are displayed Select the Distance type to be used for the clustering procedure The drop down list allows you to choose from the following Euclidean distance This is the most usual n
74. dialog in General View 213 4 4 x 2D scatter plot 184 A activation key 6 13 21 301 302 active cell 31 Add Design Variable dialog 104 Add Non design Variables dialog 107 add plot 134 add start noise 260 262 271 273 additive noise 303 Adjust to Simplex button 107 108 adjust width 133 analysis of effects 278 Analysis of Effects dialog 278 Analysis of unfolded 3 D data 81 Analysis of Variance dialog 202 ANOVA 202 278 ANOVA table 202 APC files 61 append 125 category variable 126 mixture variables 126 samples 125 variables 125 Append Samples dialog 125 Append Variables dialog 125 area normalization 222 223 ASCII file flat 53 flat wide 53 54 ASCII files export 72 import 53 import 3 D 68 ASCII MOD 75 autoscale 154 axial design extend 103 B back to suggested PC 165 Background Information Setup dialog 227 baseline 231 BFF3 files 64 bitmaps 92 93 Box Behnken design 98 99 Brimrose files 64 build design from scratch 98 button back to suggested PC source 165 calibration source 163 explained variance source 170 next horizontal PC source 168 next vertical PC source 164 previous horizontal PC source 167 previous vertical PC source 163 residual variance source 171 studentized residuals source 171 validation source 163 X1 variables source 170 X2 variables source 170 X variables source 170 Y residuals source 171 Y variables source 170 C calibration 163 ca
75. double clicking on the item in the list Tune your level names lower and upper limits until you get the desired specification Once you have specified each variable you may click OK The selected variables will be converted accordingly Note After conversion you may use Modify Properties in order to change the level names or add new levels However this will have no retroactive impact on level definition as related to the original continuous variable Warning The Edit Convert to Category Variable command replaces the original continuous variable with the resulting Category variable If you want to keep your original continuous variable in the table use Copy and Paste before the conversion 130 e The Unscrambler Menu Options The Unscrambler Program Operation Edit Convert to Mixture Variable This function converts a set of non designed continuous variables to mixture variables The function is available only for non designed data tables A dialog opens where you can select a set of variables to convert e g the Selected Variables and you can change the MixSum value before the variables are converted Note1 After converting your variables you can use Edit Correct Mixture Components so that all the mixture variables in your data table the old ones as well as the newly converted ones give a sum per experiment equal to MixSum Note2 The Convert to Mixture Variables option is disabled for all 3 D data tables Edit
76. has about the same shape as for a factorial design see Add Design Variable Dialog Factorial or Plackett Burman Design but in addition the Low Star Center and High star levels are displayed Dialog Add Design Variable for a Central Composite design Add Design Variable ID A Name RH Data Type Cancel e Continuous C Help Levels Low Loy i High Star Cube Gube Star 0 31 oiae Ain 2 08 peca Star points are used in Central Composite designs only The star point values are calculated automatically depending on the number of design variables After you have defined all your design variables you may change the star distance to center in the Define Design Variables dialog This is useful when a star point is placed in an invalid area e g negative concentration Using another distance than the default distance however changes the statistical properties of the design Be careful not to change that distance if you do not have to 106 e The Unscrambler Menu Options The Unscrambler Program Operation Add Design Variable Dialog Mixture Variable Give the new mixture variable a name and specify its lower and upper bounds Press OK Dialog Add Design Variable for Mixture variables Add Design Variable ES ID E Name Mik Data Type een Cancel inure j L m Levels e Lower Bound fi 0 0 Upper Bound 90 q Add Design Variable Process Variable The shape of this d
77. have the following names e General e Notes e Log e Instrument Transformations e Model e Constraints All sheets are not displayed for every file Properties Dialog General Sheet This sheet displays general information on the current Unscrambler file such as the type of data location and size of the file You will also see who created the file and at which date as well as a short summary of the contents of the file 84 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog Properties General Sheet Paper Properties x General Notes Log Instrument Constraints SS Paper Type Non design data Location O TMP Walerie DEMOS Paper Size 72196 bytes Created 17 september 2002 16 08 Software Version Y7 8 Creator Val rie Lengard ML Access Read Write Summary Data Version 1 Number of Samples 118 Number of Variables Cancel Apply Help Properties Dialog Notes Sheet This sheet is where you can add comments about the data or results stored in a file their purpose the analyses you are doing etc The notes you type in follow the data file They are also copied to all result files based on the data table if you have selected this option in System Setup You do not need to enter notes regarding editing and transformations done from the menus in The Unscrambler since this is
78. importing program has to know how many values to read before continuing to the next sample Flat Wide ASCII files on the other hand have a paragraph character to denote the end of each sample Sample names and variable names can also be exported if you choose the format Flat Wide ASCII files They will be placed in the first column and in the first row respectively Decide whether text and numbers should occur within single or double quotes or without quotes in the fields Name Qualifier and Numeric Qualifier respectively It is recommended to mark text with quotes and not mark numbers because it makes it easier for importing programs to distinguish text and numbers Numbers may be delimited by different characters in different ASCII files Specify which delimiter to use in the file you want to export in the Item Delimiters field Finally specify how you want to represent missing data in the ASCII file you want to export in the field String Representation The default option is m Export JCAMP DX Files The JCAMP DX format is read by many instrument softwares This file format req uires that the X part of the data have numerical names e g wavelengths Select the sets that make up the data matrix you want to write to the JCAMP Dx file in the Sets field Define new sets in the Set Editor if the already existing sets do not contain the correct data Then in the field File Info you enter information related to the JCAMP DxX file
79. in the analysis PCA Dialog Variables Sheet Choose which variables to use in your calculation in the Variable Set field You can either choose an existing Variable Set from the drop down list or define a new in the Set Editor which you enter by clicking Define In the Keep Out of Calculation field you enter any variables which you do not want to include in the analyses indicating ranges by hyphens and separating different ranges by commas for example 1 5 7 10 13 27 This does not delete the variables from the set it merely makes a model without them Press Select to enter an Editor where you can select the variables to omit Result matrices include entries for the omitted variables but the value for these entries will be defined as missing Note Invalid variable range in the Keep Out field will result in an error message as shown below The Unscrambler Program Operation The Task Menu e 259 AS The Unscrambler User Manual The Input Yalue Is not Accepted Change Value to OK You can also choose to include interaction and square effects in the calculation by pressing the button I amp S which launches the Interaction and Square Effects Selection dialog You also need to define the Weights for your variables in the Set Weights dialog The weighting determines the role of each of the variables you can change their influence on the model Consider weighting 1 SDev if you want all variables to have equal c
80. instrument All necessary parameters for export to NSAS are not always found in the data and or model files and must be entered by you The NSAS Export dialog contains three sheets General Instrument Parameters and Math NSAS Export dialog General Sheet NSAS Export E General Instrument Parameters Math OK Lok Name Raspbeny Browse Cancel PCs E a E Y variable 110 As Comments The Unscrambler model on raspberry jams data NSAS Export Dialog General Sheet Specify the file name for the NSAS model file in the Name field Use the Browse button if you want to change the folder in which the file should be stored Then select the number of PCs you want to use in the model in the PCs field This is usually the optimal number of PCs In the Y Variable field select which Y variable you want to export a model for NSAS does only allow one Y variable at a time The number at the right end of the field shows which variable number it had in the Y variable Set The Comments field allows you to write comments you want included in the model file NSAS Export Dialog Instrument Parameters Sheet The instrument parameters which will be written to the NSAS file are displayed on this sheet Most of the parameters are collected from the data and or model file and cannot be changed 76 e The Unscrambler Menu Options The Unscrambler Program Operation NSAS Export Dialog Math Sheet This sheet is used to specify whic
81. is a comma by default You can also use the clipboard copy and paste to transfer data to a spreadsheet but this will only transfer the data not the names 40 e Communication with Other Software and Instruments The Unscrambler Program Operation Exporting a Model in ASCII MOD Format All results necessary to define a model can be written in the ASCII MOD format allowing another program to read in everything necessary to reproduce the model you have made in The Unscrambler You can also export the model to NSAS or Vision format etc This option allows you to export an Unscrambler model to another program using the ASCII MOD file format Go to File Export Model and select the kind of result file you want to export Type in the name of the file you want to export in the Export dialog or click the Browse button to get a list of all available models The ASCII MOD file structure is described in the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices Exporting a Result Matrix You can also export any result matrix to another file format e g as an ASCII file A data table can be written from an Editor to disk in different file formats e g JCAMP DX a file format used by many spectroscopic software programs Plain ASCII file format is the most commonly used as it is easily read by most other programs You must import result matrices into an Editor usin
82. is missing m The Unscrambler Program Operation Glossary of Terms e 309 Replicate Replicates are experiments that are carried out several times The purpose of including replicates in a data table is to estimate the experimental error Replicates should not be confused with repeated measurements which give information about measurement error Response Variable Observed or measured parameter which a regression model tries to predict Responses are usually denoted Y variables Responses See Response Variable RMSEC Root Mean Square Error of Calibration A measurement of the average difference between predicted and measured response values at the calibration stage RMSEC can be interpreted as the average modeling error expressed in the same units as the original response values RMSED Root Mean Square Error of Deviations A measurement of the average difference between the abscissa and ordinate values of data points in any 2D scatter plot RMSEP Root Mean Square Error of Prediction A measurement of the average difference between predicted and measured response values at the prediction or validation stage RMSEP can be interpreted as the average prediction error expressed in the same units as the original response values Sample Object or individual on which data values are collected and which builds up a row in a data table In experimental design each separate experiment is a sample Scaling
83. levels are also included in the design Leverage Correction A quick method to simulate model validation without performing any actual predictions It is based on the assumption that samples with a higher leverage will be more difficult to predict accurately than more central samples Thus a validation residual variance is computed from the calibration sample residuals using a correction factor which increases with the sample leverage Note For MLR leverage correction is strictly equivalent to full cross validation For other methods leverage correction should only be used as a quick and dirty method for a first calibration and a proper validation method should be employed later on to estimate the optimal number of components correctly Limits For Outlier Warnings Leverage and Outlier limits are the threshold values set for automatic outlier detection Samples or variables that give results higher than the limits are reported as suspect in the list of outlier warnings Lower Quartile The lower quartile of an observed distribution is the variable value that splits the observations into 25 lower values and 75 higher values It can also be called 25 percentile MixSum Term used in The Unscrambler for mixture sum See Mixture Sum Mixture Sum Total proportion of a mixture which varies in a Mixture design Generally the mixture sum is equal to 100 However it can be lower than 100 if the quantity in one of the componen
84. licensee e mail ID but a different e mail ID Login as supervisor see section System Configuration in Practice and click on menu Help Change License to request a new activation key Upon submitting your new request make sure to use your registered licensee e mail ID If you do not remember what your registered licensee e mail ID is contact Camo and your consultant will either give you the information or update your registered licensee e mail ID to a new one in Camo s licensing database Symptom 4 The Unscrambler Previous Version was not Upgraded to the New Version If under starting up The Unscrambler the old version number is indicated instead of the new version number it indicates that the upgrade has not performed properly The version number of the application is also accessible in menu Help About Close the Unscrambler application un install your previous Unscrambler installation then run the new installation again Un installing your former Unscrambler version will remove the program files only all your Unscrambler data files and model files will be preserved Symptom 5 No Activation Key is Required in Standalone Installation If under starting up The Unscrambler in a standalone installation no activation key is required it indicates that the upgrade has not performed properly Close the Unscrambler application un install your previous Unscrambler installation then run the new installation again The
85. lines 199 convert individual values into levels 129 convert ranges of values into levels 129 convert to category variable 129 convert to mixture variable 131 Convert to Mixture Variable 131 Convert Vector to Data Table 80 copy 119 copy 4 plots together 149 copy all 149 copy to 292 corner sample 304 correct mixture components 129 correlation loadings 161 304 COSCIND 203 204 create new data file 45 cross validation 260 full 264 265 manual 264 265 random 264 265 systematic 112233 264 265 systematic 123123 264 265 Cross Validation Setup dialog 264 cross validation 304 cube sample 304 D Data Link Properties dialog 56 data table Non designed 131 Data table Designed 129 Non designed 89 125 126 131 database import 56 Define Design Purpose dialog 112 Define Design Variables dialog 97 104 Define Mixture Variables dialog 107 Define Model dialog 111 Define Non Design Variables dialog 107 Define Process Variables dialog 109 define sets 246 delete 126 delete item 144 derivatives 229 gap 305 gap segment 229 Norris gap 229 Savitzky Golay differentiation 230 segment 311 design 97 Box Behnken 98 99 build from scratch 98 316 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual central composite design 98 99 create new 97 D Optimal Non Mixture 98 99 extend 97 extend existing 98 fractional factorial 98 full factorial 98 Mixture 98 99 modify 97 modi
86. of X variables used to build the model Specify how many PCs to use in the prediction with the spin button Press Variance to inspect the calibration and validation variances of the selected model This takes you to the Variance Dialog where the variances are shown for each PC in the selected model Toggle between residual and explained variance by pressing the button Use the Pretreat button to pretreat the model this takes you to the Select transformations Dialog where you may choose one or several of the pretreatments that were applied to the calibration data prior to building the regression model your predictions will be based upon For an explanation of the abbreviations referring to various transformations see the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler A ppendices Finally define whether warnings should be issued by ticking or un ticking the Issue Warnings box Ticking it will enable the Warning Limits button which launches the Warning Limits dialog Prediction Three Way PLS Dialog The Prediction Three Way PLS dialog contains four sheets Samples Pri X vars Sec X vars Y reference as well as a Model field at the bottom 284 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Prediction Three Way PLS Prediction Three Way PLS Samples Pri X vars Sec X vars Y reference Sample Set All Samples 32
87. of the dialog lets you specify which Sample and Variable Sets to use on the Samples Pri X vars Sec X vars and Y Variables sheets respectively The bottom of the dialog contains general options concerning the analysis validation method number of PCs model centering Regression Three Way PLS Dialog Samples Sheet Choose which samples to use in your calculation in the Sample Set field You can either choose an existing Sample Set from the drop down list or define a new in the Set Editor which you access by pressing Define Then specify any samples which you do not want to include in the analyses in the Keep Out of Calculation field indicating ranges by hyphens and separating different ranges by commas for example 1 5 7 10 13 27 You may also select samples to be kept out by clicking Select The next field allows you to specify how to Match Samples in X and Y Data Tables e By row numbers the nth row of the 3D X data table and the nth row of the 2D Y data table are matched as one sample This option requires that samples should be presented in the same order in both tables e By sample names The software matches the nth row of the 3D X data table with whichever row of the Y data table sharing the same sample name This option allows for different sample randomisations in the two tables but requires that each sample has a unique name Note the Frozen Calibration Samples option is disabled for tri PLS Regression Three Way PLS
88. one half of the window space 6 Four sub views the Viewer window is divided into four sub views numbers 4 to 7 each of which uses exactly one quarter of the window space Select which sub view you want to copy to by choosing a number The relationship between the numbers and the sub view layout is shown below Relationship between numbers and sub view layout Sub view Layout Sub view Layout Note Go To does not erase the contents of the destination window The Unscrambler Program Operation The Window Menu e 293 Window Full Screen Sl Enlarges the current plot to fill the entire screen Clicking inside the current plot while holding down the lt Shift gt key has the same effect You can return to normal view by clicking the left mouse button or by pressing lt ESC gt or lt Enter gt Note1 For table plots numeric plots Full Screen can be activated by using SHIFT click on the area containing the plot header at the top of the plot and de activated by clicking on the same area Note2 Use of Full Screen and Copy To 1 e Full Screen blows up the plot to the full size of your physical screen no matter which space your Viewer window occupies This is a transitory state browser info and object info are NOT available e Copy To 1 copies the plot to the one sub view system the new size of the plot will be that of the Viewer window This is the largest you can achieve while retaining the browser in
89. option Modify Swap Samples amp Variables Pri Variable amp Sec Variable Modify Swap Samples amp Variables Pri Variable amp Sample Modify Swap Samples amp Variables Sec Variable amp Sample Modify Swap Samples amp Variables Pri Sample amp Sec Sample Modify Swap Samples amp Variables Pri Sample amp Variable Modify Swap Samples amp Variables Sec Sample amp Variable Modify Swap Samples amp Variables Pri Variable amp Sec Variable Use this option to exchange Primary variables and Secondary variables in a 3 D table with layout OV As a result e Former Primary variables are now Secondary variables e Former Secondary variables are now Primary variables e Variable names are updated accordingly e Former Samples are still Samples The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z Modify Swap Samples amp Variables Pri Variable amp Sample Use this option to exchange Primary variables and Samples in a 3 D table with layout OV As a result e Former Primary variables are now Samples e Former Samples are now Primary variables e Variable names and Sample names are updated accordingly e Former Secondary variables are still Secondary variables The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z 238 e The Unscrambler Menu Options The Unscrambler Program Operation Modify Swap Samples amp Variables Sec V
90. option Kill subwindows on close if you want all sub windows to close automatically when you close the parent window In the field Most recently used file list you can define how many of the projects you have previously worked on are to appear on the list at the bottom of the File menu Use the Empty button to remove all current entries in the file list This may be appropriate if you have files in the list that reside on other PCs in the network as this may slow the performance down In the last field Analysis you may tick one or both of two available tick boxes e The tick box Copy data notes to result notes lets you automatically copy all the notes which you take at any stage of your analysis to all result files you create from models derived from the data table The notes are originally stored in a file which follows your data table e The box Reduce result file size no residuals lets you select not to save all matrices in a result file when you close the file The matrices which are not stored are Eix Fiy xWeighted yWeighted This means that before you save the file for the first time or before you close it you can view all types of plots but once you have done this you can only view plots which do not contain any of the matrices not saved The Unscrambler Program Operation The File Menu e 91 The Unscrambler User Manual System Setup Dialog Editor Sheet The options provided on this sheet apply only to the Editor Stage
91. original viewpoint using the Viewpoint Reset command View Viewpoint Change The viewpoint is changed in the Change Viewpoint dialog In the Up Angle a field you enter the angle in degrees from which you want to view the plot in the vertical direction 0 degrees is the XY plane Then you enter the angle from which you want to view the plot in the horizontal direction in the field Right Angle b 0 degrees is along the XZ plane 156 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Change Viewpoint Change Viewpoint x Up Angle a el Right Angle b 5 5 You may return to the original viewpoint with the Viewpoint Reset command View Viewpoint Reset Use this command to return to the original viewpoint after you have changed the viewpoint using the Viewpoint Change command View Rotate 3 This function allows you to rotate landscape response surface plots and any other 3 D plot Use this feature in order to e get a better overview of the general shape of your response surface see figure below e study the spatial distribution of your variables and samples You can easily spot groups outliers or even errors in your data This function is accessed from the viewer menu View Rotate or from the specific icon shown above Use either the left mouse button or the keyboard arrows to perform the rotation This function provides a vertical rotation of 90 CT and 4 arrows
92. others A sample used in Mixture or D optimal designs located at a corner Centroid A sample used in Mixture or D optimal designs located at the center of an edge or face 242 e The Unscrambler Menu Options The Unscrambler Program Operation Interior A sample used in Mixture designs located inside the simplex Reference A sample with response data only The design variables have missing data Variable X Properties In the General field in the Variable Properties dialog you can modify the name of the active variable in the Name field and see whether it is a design variable or a non design variable in the Type field The Data field displays the data type category or continuous Dialog Variable Properties Variable 1 Properties E General en ER Name stirring Rate Cancel Type Non design cca Data Continuous Help m Levels Level Value Add Delete ee The field Levels displays the levels of the chosen variable The number of levels in designed data tables is fixed and cannot be changed Hence the Add and Delete buttons will be disabled You can however change the levels of the variable by selecting the level you wish to change and typing in the new value in the field below the list Click once in the list to update it Note Do not change the levels of continuous variables unless you have a very good reason to do it For category variables you change th
93. outline This option is only available when a design file is marked in the file dialog Initially you enter the dialog Print Lab Report If you want to include your own notes in the print out select the Notes check button If you want to change the page setup select the Setup button which brings you to the Print Setup dialog When you are satisfied with your choices click OK File Print ei Ctrl P You enter the Print dialog which lets you control how you transfer the current document to a printer or file You may print either Current plot or All Plots Select Current Plot if you wish to print out only the currently active plot on screen select All Plots if you whish to print out all plots currently shown on screen In the field Print range you decide what to print by selecting the appropriate radio button The print range applies to the current window in the Workspace Use Selection if you have selected a range in the current window which you want to print Finally use the spin button to select how many copies of your document you want to print Note You need to have a file open in the Editor or the Viewer to have access to this option 82 e The Unscrambler Menu Options The Unscrambler Program Operation Printing out Several Plots The Print dialog for plots offers the possibility to print either Current plot or All Plots Select Current Plot if you wish to print out only the currently active plot on screen select A
94. plotting transformation and analysis A set is a vital part of the philosophy behind handling data with The Unscrambler Dialog Set Editor Set Editor There are two categories of sets for each vector orientation samples variables The predefined sets are maintained automatically by the program and can not be modified or removed by the user The userdefined sets however can be freely manipulated Show sets of type Userdefined E Variable Sets Name Userdefined Size Interval x Spectra 16 4 19 Y Non spectra 31 3 The very first thing you need to do in the Set Editor is to select whether you want to specify a new Sample or Variable Set This is done from the drop down list in the field Show sets of type Note Always make sure that you are displaying the correct set type i e that you have made the correct choice in the Show sets of type field Depending on your choice of set type i e Sample or Variable Sets the list below will display all current sets defined by you or any other User You can delete an already defined set by selecting it on the list and pressing Remove In addition The Unscrambler automatically predefines certain sets such as All Samples and All Variables These cannot be altered but you may view their contents by selecting Predefined from the Show sets of type field Define new sets by pressing the Add button This will launch the New Sample Set New Variable Set dialog respective
95. registered all your UDIs click Cancel to exit Note Once a new UDI has been registered it can be removed later on with the command File UDI Unregister UDI Supervisor only 70 e The Unscrambler Menu Options The Unscrambler Program Operation File UDI Unregister UDI This command launches the UnRegister UDI dialog where you can remove a previously registered UDI File Export Di This is a menu option which lets you export selected parts of a data matrix to an external file in the format you define How to Export When you select File Export you enter the Export dialog which is similar to the Open File dialog see File Open Here you must specify which file you want to export and select the target format from the drop down list Export as type Then you press Export upon which you will enter a dialog which will differ depending on what your chosen target format is Export 3 D If you select a 3 D data table as source for an export the result will have a 2 D matrix structure even if the target for the export supports 3 way data Example Upon exporting data from a 3 D data table to Matlab 5 1 you will get a two dimensional array as a result More on How to Export Chapter File Export Model gives further information about how models made with The Unscrambler can be exported The Unscrambler supports OLE Drag n Drop which may be a better way to export data to another Windows program See chap
96. role as described in the Details sections hereafter Once you have chosen whether to combine Rows or Columns select the matrix or matrices you want to import from the list displayed in the middle of the dialog box If a selected matrix has several planes remember to fill in the relevant PC number s in the PCs field or the desired segments using the Segments field below the list You may enter either one value e g 4 meaning you want only PC4 several values separated by e g 0 4 meaning you want PCO and PC4 or a range e g 1 4 meaning you want PCs to 4 The Unscrambler Program Operation The File Menu e 67 The last choice you have to make is Names If you tick the box Insert row and column names the names stored in the source matrix will be included in the new table Tick the box Add matrix name to row name when combining Rows or Add matrix name to column name when combining Columns to get non ambiguous names in the case of several source matrices Click OK to access the next dialog Import 3 D where you will specify your 3 D table layout see The Import 3 D Dialog for a detailed description Importing several matrices into the same data table If you combine Rows the first selected matrix determines the number of columns in the resulting table All matrices with a different number of columns are grayed out you may select additional matrices to be imported simultaneously among the remainin
97. samples which are far from the center within the space described by the model If a sample has a very large leverage it may be different from the rest and can be considered to be an outlier Large leverage indicates a high influence on the model In the Leverage dialog choose a Line or a Matrix plot in the Plot field The Line plot presents the leverages for each sample while the Matrix plot shows the leverages for different samples and components If you select a Line plot you must also choose the number of components in the Components field and if the Samples field is enabled whether you want to see the calibration and or validation samples Plot MCR Overview Accessed with the command Plot MCR Overview this command produces a quadruple plot of the main results of a Multivariate Curve Resolution MCR These are Estimated Concentrations Estimated Spectra Sample Residuals MCR Fitting and Total Residuals MCR Fitting as shown in the Plot field of the MCR Overview dialog Dialog MCR Overview MCR Overview xj Cancel Help m Plot Estimated Concentrations Estimated Spectra Sample Residuals MCR Fitting Total Residuals MCR Fitting Components Suggested no of components is 4 4 In the Components field you may tune up or down the number of components for which the results will be plotted 190 e The Unscrambler Menu Options The Unscrambler Program Operation Plot Estimated Concentr
98. sheet for 2D Scatter 3D Scatter or Normal Probability plot Options General d OK M Equal Scale Cancel oct Help Vertical Line I Connection Line M Plot Header Markers Layout Name AACA Td a N O D O O O O O Number C Symbol Variable Level Indication Name Position on File Options in the General sheet apply to all the data points in the plot The check boxes in the upper field are used to characterize the plots more closely and are enabled and disabled according to which type of plot is activated The available boxes and their characteristics are listed below General options for 2D Scatter 3D Scatter or Normal Probability plots Option Description Grid Displays a grid allowing you to see the data points coordinates more clearly Equal Scale Applies equal scaling along all axes in the plot Curve ID Identifies each curve in the plot with a name or number Vertical Line Draws vertical lines from the data points to the horizontal plane in 3D scatter plots Connection Line Draws a line between consecutive data points This is useful to see the development when the data points are sorted Plot Header Draws the title of the plot and axis names In the Markers Layout field you choose how markers and labels will appear in the plot Available options are given below Markers layout options Options Description elect which part of the name to use in the plot with the ti
99. that define the matrix you want to normalize If necessary select Define to enter the Set Editor and define new sets Then select the normalization type in the Type field These 6 normalization methods are available 1 Area normalization Area normalization attempts to correct the spectra for indeterminate path length when there is no way of measuring it or isolating a band of a constant constituent The transformation normalizes a spectrum X by calculating the area under the curve for the spectrum 2 Unit vector normalization Unit vector normalization normalizes sample wise data X to unit vectors It can be used for pattern normalization which is useful for pre processing in some pattern recognition applications 3 Mean normalization Mean normalization is the standard normalization that is used within chromatography The areas below the spectra are made equal The results of a Mean normalization on two different samples are listed below Result of a Mean normalization on two different samples Before fter i 2 3 4 04 08 12 16 2 4 6 8 D I 08 12 16 4 Maximum normalization Maximum normalization is a normalization that polarizes the spectra The peaks of all spectra with positive values touch 1 while spectra with values of both signs touch 1 5 Range normalization Range normalization involves scaling all samples to a common range for example between 0 and 1 Thus each axis in a plot of range scaled
100. the analysis so that you can study how it correlates to the other variables by plotting Correlation Loadings Variables which are not passified may be called active variables Passify New weighting option which allows you by giving a variable a very low weight in a PCA PCR or PLS model to remove its influence on the model while still showing how it correlates to other variables PCs See Principal Component Percentile The X percentile of an observed distribution is the variable value that splits the observations into X lower values and 100 X higher values Quartiles and median are percentiles The percentiles are displayed using a box plot PLS1 Version of the PLS method with only one Y variable PLS2 Version of the PLS method in which several Y variables are modeled simultaneously thus taking advantage of possible correlations or collinearity between Y variables Prediction Computing response values from predictor values using a regression model To make predictions you need e a regression model PCR or PLS calibrated on X and Y data 308 e Glossary of Terms The Unscrambler Program Operation e new X data collected on samples which should be similar to the ones used for calibration The new X values are fed into the model equation which uses the regression coefficients and predicted Y values are computed Primary Sample In a 3 D data table with layout O V this is the major Sam
101. the axes are Principal Components or PLS components e g Scores Loadings or similar e Line plot or 2D Scatter plot displayed for a specific number of PCs e g Influence plot Regression coefficients Predicted vs Measured or similar Note The command is not available for 3D Scatter plots Example Regression Coefficients plot from a PLS regression model displayed for a model with three latent variables PCs gl EE s ei nok oad abn nynad IERE aes 2 0 F NEE X variables xvard xvard Zar xvar10 varia xvarl 6 Milk CyY var PC Milk Level 3 BOW 5 863806 After using the Next Horizontal PC button the Regression Coefficients plot is displayed for a model with four PCs Regression Coefficients 8 X variables var ward Xvar ar 0 Xvar13 var 6 Milk Grat PC Milk Level A BOW 5 323990 The Unscrambler Program Operation The View Menu e 169 See also Previous Vertical PC Next Vertical PC Back to Suggested PC and Previous Horizontal PC Note To access the Previous Vertical PC Next Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC buttons activate the Source toolbar View Toolbars View Source X Variables x Shows or hides the X variables in a plot based on variables This option is used in connection with loadings variances and variable plots You cannot disable both X and Y variables View Source X1 Variables xt
102. the lab report for the design Press Close to exit the dialog without viewing any file Results PCA This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type PCA Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results Regression This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type Regression Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results Response Surface This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type Response Surface Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results Analysis of Effects This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type Analysis of Effects 290 e The Unscrambler Menu Options The Unscrambler Program Ope
103. the number of experimental runs it will choose the design with the highest possible resolution The first field of the dialog Number of experiments to run displays how many experimental runs the chosen design will require excluding replicates and possible center and reference samples which you may define later on The field Resolution displays the resolution of the selected design Change the resolution by clicking the appropriate radio button In the field Confounding Pattern the confoundings are shown for the current design as a list Effects which cannot be studied independently from each other are displayed on the same line of the list For example A BC DE means that the main effect or A is confounded with the interaction between B and C and the interaction between D and E Effects up to three variable interactions are displayed on the list However you should not worry too much about three variable interactions since they are usually negligible What matters most is whenever main effects and two variable interactions are confounded with each other The generators of the design are shown in the Information field Details on Resolution and Confoundings The confoundings for the various resolutions available in the dialog Design Type are shown below Resolutions and confoundings for various types of designs Resolution Confoundings Resolution III Main effects are not confounded with each other Main effects are
104. the registration attempt succeeded or failed Select Close to return to the User Defined Transformation dialog Warning Trying to register a component may potentially cause The Unscrambler or your computer to hang or crash This is beyond the control of The Unscrambler and most likely due to the component itself or because of an inconsistency in the underlying Registry database You should make sure you have saved your other work before attempting to register so you do not run the risk of losing important data in case of a crash User defined 3 D User defined transformations are disabled on 3 D data Modify Sort Samples With the help of this option you can sort your samples by the names or values in one or two variables You cannot sort samples with locked cells e g designed data All defined sets are kept during the sorting Dialog Sort Samples Sort Samples E pes Samples Cancel wen 1 Help Variables g ariable 5 z Deme Sort by Order Names Ascending Values C Descending m Keys tst Mat IR a Sepal Length 2nd Mat fe Sepal Lenath 234 e The Unscrambler Menu Options The Unscrambler Program Operation You define how you want to sort the samples in the Sort Samples dialog First select the sample and variable sets that you want to sort in the Scope field If necessary press Define to enter the Set Editor and define new sets In the Sort by field you select w
105. to get back to the original layout The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z Note Only the last operation can be undone If you have already applied CTRL 3 several times continue until you get back to the original layout Modify Undo cl or Ctrl Z This option reverses the last operation you performed on your data in the Editor The following operations can be undone From the Edit menu e Cut e Paste e Fill e Insert 240 e The Unscrambler Menu Options The Unscrambler Program Operation Append Delete Convert to Category variable Correct Mixture Components Convert to Mixture Variables From the Modify menu An undone operation can be re done with Modify Redo Compute Transform Sort Samples Sort Samples by Sets Sort Variables by Sets Shift Variables Reverse Sample order Reverse Variable order Swap 3 D Layout Swap Samples amp Variables Toggle 3 D Layouts Note1 You can only undo an editing operation if it is the most recent operation you have performed Note2 To access the Undo and Redo buttons activate the Main Editor Toolbar use menu View Toolbars from the Editor Modify Redo BS or Ctrl Y You can recover the results of an editing operation you have just undone see chapter Modify Undo with the help of the Redo command The following operations can be re done From the Edit menu Cut Paste Fill Insert Appe
106. to switch to the next PC along the vertical axis of the active plot Note You may also use the lt arrow up gt key as keyboard shortcut for this command The command is available in the Viewer for PCA PCR MCR PLS1 PLS2 and three way PLS results on 2D Scatter plots where the axes are Principal Components or PLS components Note 164 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS The command is not available for Line plots or 3D Scatter plots Example Scores plot from a PCA model displayed for dimensions PC1 against PC2 4 PC2 Scores C4H3 DAD 3 H2 CS 0 2 ek DCH Jams unscaled P X expk 58 28 After using the Next Vertical PC button the scores plot is displayed for dimensions PC1 against DCH ec Jams unscaled P X exphk 56 11 See also Previous Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC Note To access the Previous Vertical PC Next Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC buttons activate the Source toolbar View Toolbars View Source Back to Suggested PC ES The Unscrambler Program Operation The View Menu e 165 Camo Software AS The Unscrambler User Manual The command View Source Back to Suggested PC and its button Back to Suggested PC is part of the PC navigation tool
107. to the database The Unscrambler Program Operation The File Menu e 57 Camo Software AS The Unscrambler User Manual Dialog Data Link Properties Connection sheet EN Data Link Properties Provider Connection Advanced Al Specify the following to connect to SOL Server data 1 Select or enter a server name CAMO SERVER D Refresh 2 Enter information to log on to the server C Use Windows NT Integrated security Use a specific user name and password User name sr Password I Blank password Allow saving password 3 Select the database on the server pubs Attach a database file as a database name pub Using the filename OK Cancel Help 0K Gel He Specify the three fields 1 Select or enter a server name this combo box lists all servers available on your network 2 Enter information to log on to the server specify the access mode either through Windows NT authentication or with a user name and password Tick off Allow saving password if you wish to save the connection details 3 Select the database on the server this combo box lists all databases available on the selected server Choose the database from which you want to import Once everything is specified hit Test Connection to check whether you have successfully established contact with the desired database Click OK to access the next dialog SQL Query Dialog SQL Query Dialog The SQL Query Dialog
108. use to open sub chapters or move up or down forward or back with the top buttons Read more about how to navigate and operate the Help system in the details hereafter The Unscrambler Help Toolbar The Unscrambler Help toolbar T l eise F Hide Previous Next Back Forward Print Options 1 The Hide button hides the leftmost window then only the current Help page is displayed When the leftmost window is hidden the Hide button is replaced by Show hit this if you want the leftmost window to be displayed 2 The Previous and Next buttons apply to pages that are located just before Previous or just after Next the currently displayed page in the Help book 3 Back takes you one step back in the path that you have followed when reading the Help system For instance if you follow a link to a new page hit Back to go back to the page where you clicked on the link You may then hit Forward if you wish to display the linked page once again etc 4 The Print button brings up a dialog where you can choose between printing out e only the current chapter displayed in the main Help window e or the chapter which is currently selected in the table of contents all its sub chapters The Contents Index and Search Sheets The leftmost window consists of three sheets Contents Index Search The tabs let you select which sheet to display The Unscrambler Program Operation The Help Menu e 297 Camo Software AS Contents T
109. which contains the imported three way data will be created in The Unscrambler Read more about the File Import 3 D menu options in chapter File Import 3 D p 66 How To Export Data With the File Export option you can export data from an Unscrambler data table to an external program In the Export dialog you will have to select the required export format and you will then access a new dialog where you may specify the parameters of the export Read more about the File Export menu options in chapter File Export p 71 How To Export Models With the File Export Model options you can export an Unscrambler model to an external format so that the model can be used by an instrument or another software for prediction or classification An export dialog allows you to specify the details of what to export depending on the export format Read more about the File Export Model menu options in File Export Model p 74 The Unscrambler Program Operation Import Export in Practice e 43 The Unscrambler Menu Options The chapters that follow describe the detailed program operation with the help of the system of menu options available from the Unscrambler s Main window Editor and Viewer These options are grouped into the following sub menus The Unscrambler Menu System Sub menu See description page The File Menu The Edit Menu The View Menu The Plot Menu The Modify Menu 217 The Task Menu 253 The Result
110. you can produce a double plot showing an Influence Plot and a Y residuals vs Predicted Y plot It is tailor made for easy visualization of outliers and their influence on the calibration model 188 e The Unscrambler Menu Options The Unscrambler Program Operation Choose which Y variable you want to investigate from the list in the Y variable field and in the Components field the components you want to examine for outliers You must also decide whether to look at the X Y or both variables in the Variables field Residuals Dialog Patterns Sheet On the Patterns sheet you can produce a double plot showing Normal Probability of Y residuals and Y residuals vs Score This is an excellent way to visualize possible clear patterns for instance curvature or systematic errors in the Y residuals when a regression analysis is performed Such structures can indicate a lack of model fit Choose which Y variable you want to investigate from the list in the Y variable field and in the Components field the components you want to examine for outliers You must also decide whether to take into account the calibration and or validation samples in the Samples field Residuals Dialog MCR Fitting Sheet From the MCR Fitting sheet available only for MCR results you can produce a line plot showing the Variable Sample or Total residuals which summarize the quality of the MCR fitting Dialog Residuals MCR Residuais x MCR Fitting PCA Fitting
111. your non mixture design variables The top field displays information about all defined design variables while the Information field displays information about the currently selected variables and design The dialog allows you to add a new or to delete or change an already defined variable by clicking one of the buttons New Delete or Properties Clicking New or Properties leads you to the Add Design Variable or Variable Properties dialog where you can specify the name and properties of the process variable If two or more of your variables are linked by a linear relationship check the Multi Linear constraints box and click Next Define Process Variables Next Dialog Once you have added all necessary process variables by clicking New or changed their definitions by clicking Properties you may proceed by clicking Next Depending on your choices the next dialog will be the following e If you checked the Multi Linear Constraints checkbox Enter Multi Linear Constraints e If not Define Non design Variables Design Wizard Enter Multi Linear Constraints In this dialog you can define up to 20 constraints between your process or mixture variables If a constraint falls out of the experimental region an information dialog opens warning you that you have either to change this constraint or remove it from the design specifications The Unscrambler Program Operation The Design Wizard e 109 Camo Software AS The Unscrambler User Man
112. 04 2003 12 18 S ISRT dll 320 KB Application Extension 01 04 2003 12 18 Users dll 184KB Application Extension 01 04 2003 12 18 3 objpss dll 32KB Application Extension 01 04 2003 12 19 Then double click on the setup exe again to execute it Symptom 2 Invalid Activation Key When clicking the Activate button in The Unscrambler Activation dialog the following message might appear The Unscrambler A This activation key is invalid for your computer Please contact CAMO ses This message will occur if the Unscrambler License Server activation key you have input is not valid for your machine Provide Camo with your machine code to receive the appropriate activation key for your machine see Requesting an Unscrambler Activation Key above 20 e Installation and Configuration The Unscrambler Program Operation Symptom 3 Incorrect License Details in Standalone Installation The license details information of your Unscrambler Standalone installation doesn t correspond to the type of license you have purchased from Camo For example the license appears to be a 30 day license instead of a permanent license m License Details The Unscrambler 9 6 Trial Version 24 Days remaining Click on buy for full version of The Unscrambler e e For any queries contact support camo com or Cancel support camo no This situation may occur if when requesting an activation key you have not utilised your registered
113. 102 Extend an Axial Design 103 Fill 120 Fill Missing 120 121 Gap Segment Derivatives 229 230 Gaussian Filter Smoothing 222 Generate Design 113 Get Centered Models 280 Get Model 280 Goto 131 Goto 3 D 131 GRAMS32 Import 61 Histogram Plot 176 Import 50 51 52 Import ASCII 53 Import Target 50 Import Worksheet 55 Important Variables 195 Important Variables three way PLS 195 Information 122 Interaction and Square Effects Selection 111 Interaction and Square Effects Selection 251 Item Properties 144 JCAMP DX Import 59 Last Checks 118 Layout 245 Leverage 190 Licensee Information 302 Line Plot 175 Line Properties 144 Loading Weights 185 Loadings 185 Locate User Defined Transformation 234 235 login 26 Login 26 Matrix Plot 177 MCR 267 MCR Overview 190 Median Filter Smoothing 221 Modify Primary Variable Set 250 Modify Sample Set 250 Modify Secondary Variable Set 250 modify supervisor 25 Modify Supervisor 25 The Unscrambler Program Operation Index e 317 modify user 25 Modify User 25 Modify Variable Set 250 Moving Average Smoothing 220 Multiplicative Scatter Correction 225 226 Multivariate Curve Resolution MCR 267 New Data Table 45 New Primary Variable Set 249 New Sample Set 247 New Secondary Variable Set 249 new user 25 New User 25 New Variable Set 248 Noise 228 Normal Probability Plot 176 Normalize 222 223 Norris Gap Derivatives 229 NSAS Export 76 Open Fil
114. 125 select transformations dialog 90 Select Transformations dialog 90 select variable sets to sort dialog 237 Select Variable Sets to Sort dialog 237 Select Variables dialog 133 252 sensitivity to pure components 311 Set Editor dialog 246 set up design 99 Set Weights dialog 262 sets define 246 edit 246 modify primary variable set 250 modify sample set 250 modify secondary variable set 250 modify variable set 250 new primary variable set 249 new sample set 247 new secondary variable set 249 new variable set 248 sheet range 55 shift to other window 1 2 297 variables 238 Shift Variables dialog 238 significance level choose 149 significance testing methods 203 SIMCA classification 280 smoothing 220 gaussian filter 222 median filter 221 moving average 220 Savitzky Golay 221 SNV 232 SOD 257 sort samples 235 sort samples by sets 236 Sort Samples dialog 235 236 sort variables by sets 237 source 162 spaces in ASCII import 53 54 Specify Levels dialog 124 SpectrOn files 63 Spectroscopic Transformation dialog 224 spectroscopic transformations 224 split category variable 131 SQL Query dialog 58 stability plot 160 326 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual standalone installation 1 standard sample sequence 154 standard normal variate 232 Standard Normal Variate 232 Standard Normal Variate dialog 232 standardization 314 standardization of variables 311 star p
115. 13 27 Clustering Dialog Sum Of Distances and Number of Iterations The K Means clustering algorithm consists of the following three steps 1 The algorithm is initiated by creating k different clusters The sample set included in the analysis is first randomly distributed between these k different clusters 2 As a next step the distance between each sample and its respective cluster centroid is calculated 3 Samples are then moved to a cluster k that records the shortest distance from each sample to the cluster centroid The Number of Iterations is the number of times the K Means algorithm is repeated to obtain an optimal clustering solution every time starting with random initial clusters The Unscrambler Program Operation The Task Menu e 257 The Sum of Distances SOD is described as the sum of the distance values of each sample to their respective cluster centroid summed up over all K clusters This parameter is uniquely calculated for each batch of cluster IDs resulting from one single iteration of the algorithm The results from all iterations are compared and the solution with least SOD is retained Recommended Procedure for Clustering Start the analysis with a small Number of Iterations say for example 10 if the sample set contains 500 samples After obtaining the first clustering results keep the clustering variable in the data table and launch a new analysis with a larger Number of Iterations thi
116. 2 Mixture Process 3 Process 200 e The Unscrambler Menu Options The Unscrambler Program Operation Mixture Select this plot type if you want to build a response surface with three mixture variables varying The other mixture variables and all process variables are fixed This plot is available only if your analysis included mixture variables Note This plot has a triangular shape and displays the mixture variable names and ranges of variation Mixture Process Select this plot type if you want to build a response surface with one process variable and one mixture variable varying The other mixture variables are varied so that all the coordinates of the surface lie along the selected mixture variable axis All process variables except the varied one are fixed This plot is available only if your analysis included mixture variables and at least one process variable Process Select this plot if you want to build a response surface with two independent process variables i e non mixture variables varying All the other process variables and all mixture variables are fixed This plot is available only if your analysis included two or more process variables Note The Process and Mixture type plots are able to handle multi linear constraints A border is drawn around the actual experimental region and only the points belonging to this region are drawn Configure Ranges of Variation Once you have chosen the type of plot to use
117. 299 Help Quick TOU a cscscdsstcosass E E es 300 Help got the Days eege deed ec 200 IG pA DOU betes insists ci deg egene ENEE ege SEENEN 301 Helps Chan ge License eet sarees tens snider ae ar aaa a Ea o e EE eebe Dee 301 Help LICENSEE iisen E E T EEY 302 Glossary of Terms 303 Index 315 x e Contents The Unscrambler Program Operation Installation and Configuration In this chapter you will learn about system requirements for installing and running The Unscrambler installation procedure and system configuration after installation The Unscrambler satisfies the FDA s requirements for 21 CFR Part 11 compliance System Requirements This section presents hardware and software requirements for installing and running The Unscrambler Hardware Requirements We recommend that you use at least a Pentium PC running at 100 MHz or more Memory space is an important issue at least 16 MB of RAM should be available preferably 32 MB Using a more powerful PC improves performance significantly and is advisable if your data tables are large Software Requirements The Unscrambler software is written for the Windows 95 Windows 98 Windows NT 3 51 or later Windows 2000 and Windows XP operating systems The Program does not run on Windows 3 x or Windows for Workgroups platforms Standalone Installation 1 User License Desktop Installation Installing The Unscrambler Standalone is a straightforward procedure in five steps 1
118. 5 Number of left side points 1 Number of right side points E E The number of smoothing points is 5 Polynomial order 0 E In the Scope field in the Savitzky Golay Smoothing dialog select the Sample and variable sets that define the matrix you want to smooth If necessary use the Define buttons to enter the Set Editor and define new sets You may then select the smoothing points in the Parameters field Note that a larger range will give a smoother shape to the sample but you may lose some information There are two ways of selecting the smoothing points e Either type in the Number of smoothing points it is always an odd number 2n 1 and the smoothing segment will be symmetrically distributed on both sides of each curve point n left side points n right side points e Or choose the Number of left side points and right side points separately Finally enter the order of the polynomial to be fitted in the sub field Polynomial Order For instance a polynomial order of 2 means that a second degree equation will be used to fit the data points Note that after the operation is completed the data will be slightly truncated at both ends If p is the number of left side points and q the number of right side points in the smoothing segment the first p and the last q variables in the smoothed variable set will be set to zero This is because there are not enough points to the left resp right of these variables
119. 65 File Import Unscrambler Results Ctrl U You can read result files made by earlier versions of The Unscrambler into the Editor to study their numerical values by selecting the file type and file you want to import in the Import dialog You then enter the Import from X dialog where X represents the result file type see the figure below for an example You select the matrices you want to import from the matrix list See the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler A ppendices for a list of all matrices used in The Unscrambler The Unscrambler Program Operation The File Menu e 51 Dialog Import from PCA Result Import from PCA Result Name PCA Result Combine C Rows Columns Cancel Help ol Matrix name Rows Cols Pins Names I Insert row and column names J Add matrix name to row name T Add matrix name to column name rm Current selection Number of rows 0 Number of columns 0 Total number of elements 0 First you choose whether you want to import rows or columns by clicking the appropriate radio button Then select the matrices you want to import from the file Several matrices can be imported together by combining either rows or columns Matrices that will not match the ones already chosen will be grayed out You can import several matrices either by rows or columns depending on whether
120. 7 Ce Min max e Cancel Classes _ Cancel Help Px w Nl ooo Gu Variance In the Methods field choose whether you want to mark by Min max or by Classes Then in the list below specify the number of PCs listed in the left column for which you want to mark samples and how many listed in the right column No samples are marked for PCs with 0 in the right column i e in the above figure only PC 1 is marked If you select the method Min max a number of extreme samples will be picked out for each PC according to your specification in the right column in the list below the Methods field It will be labeled of Min max and for each Min max selected two extreme samples are marked max and min value Thus setting the number to 3 will mark a total of six samples If you select the method Classes the samples will be divided into a number of classes for each PC One pair of extreme samples max and min value will be picked out for each PC according to your specification in the right column in the list below the Methods field It will be labeled of Classes and for each class two extreme samples are marked Thus setting the number to 3 will mark a total of six samples If you wish to include variance press the Variance button to launch the Variance dialog For more general information about marking lookup section Edit Mark Basic Notions Edit Mark Significant X Variables Only 4 This command is ava
121. ALC EE 157 VIEW Projected ME 158 View Plot TD iiss occ ee ec pee eS egies ease Rade Wea sith is enced 158 KAS EE 158 View Trend RE 159 View Trend Lines Regression Lime 159 View Trend Lines Target Lane E 159 View Hotelline T2 BUM psesis cscss casascseassvnsessssessactevepesasesceneeaeestavapnsse chenscessevbetsessaneonaneresene 159 View Uncertainty Test cece cess ceeeeceeeseceesacesceeecaeaeseecaesaee ceeecaseesersaseseesseeeesaeeatageas 160 View Uncertainty Test Stability Plot 0 eee ceseeseececseeeeeaecaeeeeeaecaeneesacaeeseeeaeenes 160 View Uncertainty Test Uncertainty Limits 0 0 0 cceeeeecseeecnsceeseeesaeeesseesaeeenes 161 View Correlation Loadings oe sesseesesceeeeceeeeseesceeeceesceeesaeeecescessaeeeseeceseeeaeseaeaeeaees 161 WIGW SS OUNCE EAE E EA E EE ETT 162 View Source Calibration snieni E EE a r E S E ER 163 View Source Validation seseseeeeseeseeeesetsrerseretrrtsststistststestestrst aesaeeaaeeaeceeceesaesaaeeaeenes 163 View Source Previous Vertical PC EE 163 View Source Next Vertical PC cceecceseesceecceeeenseeeceeeeeeaeeseceaecneeeaecaeeaeeaeceeeeeeseaeeateeas 164 View Source Back to Suggested DC 165 View Source Previous Horizontal PC ccceseeseeeeeeeeeneeceeeeeeeeceeeeaeeaaeeaeeeeeceesaeeeeeneenaees 167 View Source Next Horizontal PC oo cece ceesseceseceseee seeeseecenceeseeeesaeceaaessueeeessaeeeneeeeaeeeees 168 VIEW SOUTCE Ee 170 View Source X1 Variable
122. AS data file the Instrument Information block was created based on the information found in the NSAS data file reflecting any transformations applied to the data before the NSAS data file was created For this reason you should not change any default values unless you are very confident in what you are doing Any transformation applied to the raw data using The Unscrambler must be manually specified For more information about the correspondence between Unscrambler and NSAS transformations see next chapter NSAS Transformations Press the Log button on this sheet to enter the Log where you can make notes about your export NSAS Transformations When you have applied a transformation to your data using The Unscrambler prior to calibrating a model and exporting it to an NSAS Vision model file it is of vital importance that these transformations are flagged in the NSAS Vision file to be able to perform correct predictions using the model Unfortunately there is not a simple one to one correspondence between the two programs so some manual work is required by you Each NSAS Vision transformation and the corresponding Unscrambler transformation is listed below Note that a single NSAS Vision transformation often corresponds to two Unscrambler transformations applied consecutively The Unscrambler Program Operation The File Menu e 77 Important The Unscrambler always assumes parameters like segment sizes gap sizes etc to be expressed in nu
123. Cross Validation Setup Cross Validation Setup x Method Random D OK 2 Full Cross Validation Eet Random Number off Systematic 123123123 Help e Systematic 111222333 amples p Manual Seg No of No Ges Samples Samples Not Found in Any Segment m The drop down Method menu gives you five different cross validation options to choose from These are listed below 264 e The Unscrambler Menu Options The Unscrambler Program Operation Cross Validation methods Method Description Full Cross There are as many segments as there are samples one sample per segment Leave one Validation out Random You select the number of segments or the number of samples per segment The Unscrambler picks the samples at random Systematic You select the number of segments or the number of samples per segment The first sample 123123 belongs to the first segment the second sample to the second segment etc You select the number of segments or the number of samples per segment First N samples belong to the first segment next N samples to the second segment etc You select the number of segments or the number of samples per segment The Unscrambler gives a suggestion for which samples belong to which segments You can override this by changing the sample numbers Every change you want to make has to be done manually through the box below the segment list Any samples that have not been assigned to a
124. D format varies according to the type of plot The information given in the Plot ID field together with the Plot Identification window Window Identification tells you all about the data displayed in the active plot To remove the text line uncheck the option View Plot Statistics Makes a list of statistics pop up in the plot Select this option again to remove the statistics from the plot This option is available for two types of plots only 2D scatter plots and histograms Which statistics are displayed depends on which plot you are in An overview of which options are available in which situations is displayed below Available plot statistics in various contexts Statistic Interpretation bp scatter plot Histogram Elements Number of points X X Slope of the regression line between X abscissa and Y ordinate Offset Intercept of the regression line X Correlation Correlation between X and Y X RMSEP Root Mean Square Error of Prediction X Validation RMSEC Root Mean Square Error of Calibration X Calibration RMSED Root Mean Square Error of Deviations na Standard Error of Prediction Pe SEC Standard Error of Calibration X Calibration SED Standard Error of Deviations X other 158 e The Unscrambler Menu Options The Unscrambler Program Operation Bias Average difference between Y and X amo Software AS View Trend Lines Skewness Asymmetry X
125. DI from a list then access the Import dialog Dialog User Defined Import UDI User Defined Import UDI x r Select UDI Import Cancel Help r UDI Info Location C UDI Testing JCampUDI dil Description JCamp User Defined Import DLL ls System UDI No Select the required UDI from the list then look up the UDI Info field to check whether this UDI indeed refers to the Import format you are looking for If OK click Import to access the Import dialog If the UDI you are looking for does not appear on the list check with your system administrator whether it has been registered in The Unscrambler see Related Topics below Note Clicking Import starts the User Defined Import procedure For obvious reasons The Unscrambler Help system does not cover that stage of the import Should you encounter any trouble while defining or performing the import contact your system administrator or the person responsible for that particular UDI The Unscrambler Program Operation The File Menu e 65 Camo Software AS The Unscrambler User Manual Related Topics Register a new UDI File UDI Register New UDI Un register an existing UDI File UDI Unregister UDI File Import 3 D The menu option File Import 3 D allows you to import 3 D data from various sources Choose the relevant type of file you enter the Import browser where you can search for and select the desired source file only files o
126. Data Files Location Setup will install the The Unscrambler data files in the following directory To install to this directory click Next To install to a different directory click Browse and select another directory If this installation is going to be accessed by several users in a network select a directory on a file server that all users can reach Also make sure that all users have write access to this Destination Folder C My Documents The Unscrambler DATA Browse InstallShield 9 The program is being installed The Unscrambler User Manual 4 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS InstallShield Wizard Setup Status The Unscrambler Setup is performing the requested operations Writing system registry values InstallShield 10 When the setup is complete click Finish InstallShield Wizard ES InstallShield Wizard Complete Setup has finished installing The Unscrambler on your computer Dance Requesting an Unscrambler Activation Key 11 Start The Unscrambler from the Start menu The Unscrambler Program Operation Standalone Installation 1 User License Desktop Installation e 5 The Unscrambler Programs E A Documents b a L Settings gt i Search gt SA Help and Support J Run Log Off valerie Windows XP Professional 12 The U
127. Define Keep Out of Calculation Model Name Wood Severity Q amp D i Number of Components p The top of the dialog lets you specify which Sample and Variable Sets to use on the Samples Pri X vars Sec X vars and Y Variables sheets respectively The bottom of the dialog allows you to locate a relevant tri PLS model and to select the proper number of components for prediction Note Automatic Pretreatment is not available in three way PLS prediction Prediction Three Way PLS Dialog Samples Sheet Choose which samples to use in your calculation in the Sample Set field and enter any samples which you do not want to include in the analyses in the Keep Out of Calculation field For more details you may lookup the description of a similar dialog in Statistics Dialog Samples Sheet p 253 Prediction Three Way PLS Dialog Pri X vars Sheet Choose which Variable Set to use as Primary X for prediction in the Variable Set field To define new variable sets click Define to go to the Set Editor If some of the Primary X variables were kept out from the model at the calibration stage they will automatically be kept out at the prediction stage and their numbers will be displayed in the grayed out field Keep Out of Calculation This occurs when you select the model to use see Prediction Three Way PLS Dialog Model The following warning will then be displayed The Unscrambler e 1 Some of the Pri X variables will auto
128. E 268 Regression Three Way PLS Dialog AA 273 Task ee 277 Task Analysis 0f D E tSssssisisisissiseiisiisesisiiiisisineseneetavirekensee otken esain oures EaSE oeur sasen r Etisk 278 ASK Classif EE 278 Get Model DIOS creeme erener e e r E a E E E RE 280 Text Warning list Dialogic cissseiecscvesesvsiessactsvecs banectuaes sncntaeebenebasees E E AE E 281 KE 282 Task ET EE 282 Prediction Three Way PLS Dialog 284 Task User defined Analysis cceeccsseceeesescneceeeseesecescsecsecessesseceeessasseseessasseceseeeaeenes 286 Task Recalculate with Marked occ ceeceeeseeseeceeeeeeesesseeeesessesseeeesecseeseeeesaeeaeeaees 287 Task Recalculate without Marked ccc ccc eecseeseceeecseeeceeecaeeecseeseceesssnesseeesaesaeeas 288 Task Recalculate with Passified Marke 288 The Unscrambler Program Operation Contents e ix Task Recalculate with Passified Unmarked ccccccccccccsssssecccee ceceececesseaeeececeseceaeseeseas 288 Task Extract Data from Marked A 288 Task Extract Data from Unmarked s sesssseessesesesereeeesetrsrererttsrssrsreststestrtessrestsesreseeseesree 288 The Results Menisin oran r S N E EE S 289 Feelt rnunana EE dE A N EEEE 289 Results Dialog enneren n NEER R AA ES NE EAN ENANS 289 Feelen PGAg 3 tistics necaptunatha ne E E E A A E EAT 290 UE EE 290 Results Response Surface 290 Results Analysis of Effects sscsssecssessssese conscseseressesessressonssessoseesnsenesssenserssas
129. If you wish to remove some samples make sure that the active plot contains marked samples If you wish to remove some variables make sure that the active plot contains marked variables Task Recalculate with Passified Marked Use this option when you have marked some X or Y variables on a Variable plot e g loadings regression coefficients and want the marked variables to be passified in the next analysis You will then enter the Regression dialog where the new weights for the variables to be passified will have been updated automatically This option is available whenever you have some marked X or Y variables or both on the currently active plot It is not available if you have some marked samples on the currently active plot Task Recalculate with Passified Unmarked Use this option when you have marked the important X and or Y variables on a Variable plot e g by using Edit Mark Significant X variables only and want the remaining variables to be passified in the next analysis You will then enter the Regression dialog where the new weights for the variables to be passified will have been updated automatically This option is available whenever you have some marked X or Y variables or both on the currently active plot It is not available if you have some marked samples on the currently active plot If the currently active plot contains only X variables the Y variables will not be affected If it contains vari
130. It allows you to switch back to the original axes after navigating with the Previous and Next options Note You may also use the lt gt key as keyboard shortcut for this command In practice here is what happens depending on the type of plot which is currently active e Ona 2D Scatter plot where the axes are Principal Components or PLS components e g Scores Loadings or similar go back to PC1 PC2 e Ona line plot or 2D Scatter plot displayed for a specific number of PCs e g Influence plot Regression coefficients Predicted vs Measured or similar go back to the number of PCs suggested as optimal by the system Example Predicted vs Measured plot from a PLS regression model displayed for a model with three latent variables PCs Predicted Y Elements 28 6 Slope 0 973280 Offset 0 104495 Correlation 0 986549 RMSEC 0 556502 SEC 0 566713 Bias 3 406e 07 10 Measured 0 2 4 6 8 10 Milk y var PC Milk Level 3 After using the Back to Suggested PC button the Predicted vs Measured plot is displayed for a model with one PC 166 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Predicted Elements Slope Offset Correlation RMSEC SEC Bias Measured Y 2 0 2 4 6 8 10 Milk Grat PC Milk Level 1 See also Previous Vertical PC Next Ver
131. Jite Jae Device z Scheme Default Greytones Save As Delete ltem Background D OK Cancel Help System Setup Dialog Missing Data Sheet The options provided on this sheet will apply to all data tables you open in the Editor 94 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog System Setup The Missing Data Sheet System Setup xj General Editor Viewer Missing Data m Options Ki e T Notify me when missing data is filled in Parameters Method Principal Component Analysis Y I Scale Data Before Estimating Missing Values Cancel Apply Help In the Options frame choose whether you want to Automatically fill in Missing Data When this option is checked any new table with some missing data opened in the Editor will have all its missing values filled in with estimated values prior to being displayed on screen If you also select Notify me when missing data is filled in a small information dialog will appear every time this happens as shown below The Unscrambler E A Missing data was successfully filled in In the Parameters frame choose the estimation Method Principal Component Analysis or Row Column Mean Analysis You may optionally Scale Data Before Estimating Missing Values by ticking the box at the bottom This is recommended if the variables you usually deal with a
132. N X angent ANINT CX Nearest integer rounding I ANH X H yperbolic tangent C S X can denote both samples and variables in this table Notel A commonly used expression is X Log X This expression generally transforms skewed variable distributions into more symmetrical ones Use a histogram plot or Variable Statistics in order to check whether the skewness was improved or deteriorated after applying the transformation Note2 You may Undo Redo operations performed in the Compute dialog with the corresponding options in the Modify menu Build Expression Dialog In the Build Expression dialog you can build your expression using the ready made functions and operators allowed in The Unscrambler 218 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog Build Expression x MIT V1i2 e V17 Display e Display Sample Variable Numbers Display Sample Varable Names Select Sample Variable _ Sample Variable Insert Element EEE Arithmetic Functions El D Li LNG SORT ET WI OT LOG EXP MIN K1 X2 MAX XLX2 m Trigonometric Functions SING cos SINH COSHX ACOSH TANG ASING ATANG ih ATAN2 X1 2 irs i iip Other Functions SIGNOS ANINT X AINT l The upper text field shows the expression as it is being built In Display choose
133. RET 152 Sample Statistics Dialog and Variable Statistics Dialog sesseessseseeeeeeserereereerrrresrsererrneee 153 View Sample Names tcs sse sessehelteessvazczs ts nE E E EEEE EY 153 View Point Names ccceccesecescececeeeeeseeeseeseceseceseeeeeeeseesueeceneeeseecseseaeeaaeeaeeeecesaeeeesneesaees 154 View X Y Variable Numbeng ce cece cesseeeeessceeseeesaeeeseeeaeeeseees caseeseeseeaeeaeeseaes 154 View Level Indiees eceeceeeccsccessceseeeeseeeaeescescceeeceeeceeecsecsaeaeseeeeseesaeseaeeeeeeeeeeeseeseeseeeeatees 154 View Standard Sample Sequence wo eee cece cesses ceeeceeeeesseeeeaessesseeae cesseeseesecaeeaeaees 154 View Experiment Sample Sequence o 0 ieee eeeeecseeeeneeeecaeesceeecaeeecneesscesssneeseeeeaesaseas 154 KA E 154 KAISER 154 Kat 154 Ka 155 View Scaling Min Max s issccsscasesstectevstcaseteswevenevsteusivsteswesetevetesess ENNEN 155 View Scaling EE 155 View Scaling Exact Min Max 0 ccc csccssesseeeceececssseeeesesssescaeseeeesasee caaeeeeseeseseseesneeags 156 View Scaling Exact Pram visssssscscssxses erevssctsnscesseneseesd EUn EEEa ts ETEA EARE SEERNE EESE nae 156 Mew Zoom IN eran eaa ea nin eE e E aE NEEE EEAS 156 Niew ZOOM OUt EE 156 View NIEWEN ee ett dee eraai ia Deet eet 156 View Viewpoint CHAN Ge sess scevssbesssevecies setae oniu r r SEEE ebe 156 View Viewpoint Reset ccc cscesseeecseeceseseesecescsecsecessessecee sessesaeeaesaseeseeecaseesaseneneas 157 NIG W RO T
134. S The Unscrambler User Manual Help Licensee The Licensee field of the Licensee Information dialog displays the Activation key which allows you to activate your Unscrambler license The activation key is uniquely associated to your license and to the machine on which The Unscrambler is installed The dialog also tells you how many users can have access to The Unscrambler i e how many you can have registered in the login list as well as the expiration date of your license Dialog Licensee Information Licensee Information E x This product is licensed for up to 1 user The license expires on November 27 2005 Cancel eae te p The Unscrambler Activation key MSKFKPFDSDZTMDCFRSFDBNTTSDPXGS For any queries contact support camo com or support camo no Contact CAMO if you need to be able to register more users 302 e The Unscrambler Menu Options The Unscrambler Program Operation Glossary of Terms 2 D Data This is the most usual data structure in The Unscrambler as opposed to 3 D data 3 D Data Data structure specific to The Unscrambler which accommodates three way arrays A 3 D data table can be created from scratch or imported from an external source then freely manipulated and re formatted Note that analyses meant for two way data structures cannot be run directly on a 3 D data table You can analyze 3 D X data together with 2 D Y data in a Three Way PLS regression model If you want to analyze your
135. Select the file you wish to view and press View to open a Viewer window where the results will be displayed To view only a filtered list of files select the appropriate type e g Regression from the drop down list in the Files of type field All files of the selected type located in the current directory will be shown If necessary click your way to another directory by using the tools at the top of the dialog box Tick the box Mine only if you want to filter out files created by other users than yourself The Unscrambler Program Operation The Results Menu e 289 A regression model can be exported directly from this dialog by pressing the Export button Click the Properties button if you want to review the model properties This also allows you to register transformations for automatic pre treatment PCA and Regression models only by accessing the Transformations sheet of the Properties dialog The Information field at the bottom of the dialog displays information about the selected file Press Print to get a printout of the information Pressing the Warnings button gives you an overview of the warnings that were issued during the given analysis displayed in the Text Warning List dialog Pressing the Variance button launches a table of the model variances displayed in the Variance dialog If you have chosen a file of type Designed Data the Warnings button is replaced with a Lab Report button allowing you to pre view and print
136. Selection Interaction and Square Effects Selection Eg Selected Effects Available Effects o Name 45 Cancel ud Help Add All Add Int Add Squ Remove All Show Effects from fi 2 Select Sub Set The Unscrambler Program Operation The Design Wizard e 111 Define Model Next Dialog Once you have added all desired effects to your model you may proceed by clicking Next Depending on your type of design the next dialog will be as follows e D optimal Non mixture or Mixture Design Details e Mixture with a simplex region with or without Process variables Define Design Purpose Design Wizard Define Design Purpose In this dialog which is specific to a Mixture Design with a simplex experimental region specify the purpose of your design by selecting the proper radio button calibration screening or optimization Design Wizard Define Design Purpose Design Wizard Define Design Purpose Your mixture region is a simplex your design can be generated using a classical Mixture algorithm depending on your purposes What is the purpose of your design e Generate different samples covering the mixture region uniformly e g calibration Optimization of mixture variables the model includes mixture interactions and squares Note An optimization design implies that square terms of mixture variables are included in the model This radio button will also be disabled if too many vari
137. Sheet This sheet lets you specify the Response variables in your tri PLS model Dialog Regression Three Way PLS Y variables sheet Samples Pri X vars Sec X vars Y variables Y Variable File BF_SFC v Browse Variable Set Severity age 2 e Define Keep Out of Calculation Ce Select Weights All 1 0 SDev Iw Include Weights Since the 3 D data table displayed in the Editor upon launching Task Regression only contains the X variables you have to get the Y variables from another file containing 2 D data Use the Browse button to find the adequate Y variable file In the Variable Set field choose which variables to use as Y variables You can either choose an existing Variable Set from the drop down list or define a new in the Set Editor which you access by pressing Define In the Keep Out of Calculation field you enter any variables which you do not want to include in the analysis indicating ranges by hyphens and separating different ranges by commas for example 1 5 7 10 13 27 You may also select variables to be kept out by clicking Select The Unscrambler Program Operation The Task Menu e 275 Weighting is done in the Weights field Only 1 SDev is available as weighting option you select it for all Y variables by ticking Include Weights Regression Three Way PLS Dialog General Options In the bottom of the dialog you can find general options for your analysis 1 Model Size 2 Number of
138. Shows or hides the X1 variables Primary X variables in a three way PLS result plot based on variables This option is used in connection with loading weights variances and variable plots You cannot disable X1 X2 and Y variables simultaneously View Source X2 Variables Eg Shows or hides the X2 variables Secondary X variables in a three way PLS result plot based on variables This option is used in connection with loading weights variances and variable plots You cannot disable X1 X2 and Y variables simultaneously View Source Y Variables leal Shows or hides the Y variables in a plot based on variables This option is used in connection with loadings variances and variable plots You cannot disable both X and Y variables View Source Explained Variance m Displays the explained variance for each PC in a cumulative way as a percentage of the initial variance By definition the explained variance for PC 0 is 0 The explained variance is expressed in percent and makes it easy to see how much of the total data variation is explained by the model 170 e The Unscrambler Menu Options The Unscrambler Program Operation View Source Residual Variance Displays the residual variance for each PC The residual variance plot is used to find the optimal number of components to use in the model It is an expression of the modeling or prediction error View Source Y residuals No Displays the raw Y residual
139. The Modify Menu e 219 The Unscrambler contains many standard transformations to be found as sub options in Modify Transform but you can also take advantage of the Modify Compute General function to transform the data with simple formulas e g logarithm Here are all the choices to be found under the Modify Transform menu option e Modify Transform Smoothing Moving Average Modify Transform Smoothing Savitzky Golay Modify Transform Smoothing Median Filter Modify Transform Smoothing Gaussian Filter Modify Transform Normalize Modify Transform Spectroscopic Transformation Modify Transform MSC e Modify Transform Noise Modify Transform Derivatives Norris Gap e Modify Transform Derivatives Gap Segment e Modify Transform Derivatives Savitzky Golay e Modify Transform Baseline e Modify Transform SNV e Modify Transform Center and Scale e Modify Transform Reduce Average e Modify Transform Transpose e Modify Transform User defined Modify Transform Smoothing Smoothing is used to reduce the noise in the data without reducing the number of variables It is a row oriented transformation that is to say the contents of a cell are likely to be influenced by its horizontal neighbors A sub menu to the Smoothing menu lets you choose between four different methods to smooth your data 1 Moving Average first finds a data value by averaging th
140. The Unscrambler User Manua Camo Software AS The Unscrambler Program Operation By CAMO Software AS www camo com This manual was produced using ComponentOne Doc To Help 2005 together with Microsoft Word Visio and Excel were used to make some of the illustrations The screen captures were taken with Paint Shop Pro Trademark Acknowledgments Doc To Help is a trademark of ComponentOne LLC Microsoft is a registered trademark and Windows 95 Windows 98 Windows NT Windows 2000 Windows ME Windows XP Excel and Word are trademarks of Microsoft Corporation PaintShop Pro is a trademark of JASC Inc Visio is a trademark of Shapeware Corporation Restrictions Information in this manual is subject to change without notice No part of the documents that build it up may be reproduced or transmitted in any form or by any means electronic or mechanical for any purpose without the express written permission of CAMO Software AS Software Version This manual is up to date for version 9 6 of The Unscrambler Document last updated on June 5 2006 Copyright 1996 2006 CAMO Software AS All rights reserved Contents Installation and Configuration 1 KR UM E Hardware Requirements Software ReguirementS ieies za e aeree E Aere R EEA PENEN EEEE EEE EEEE SEERE EER Standalone Installation 1 User License Desktop Installation 0 eee esse ceeeeenee ceeeeeeeeeeeeneeeaeeneees 1 Downloading The Unscramb
141. To Select Reasonable Constraints Is the Mixture Region a Simplex Models for Mixture Variables The Mixture Response Surface Plot Display The page you select from the list is displayed in the main Help window Help Button Di Gives you help about any component of The Unscrambler To use the Help button click on it this will activate the button and you will see the shape of the cursor change to an arrow with a question mark 298 e The Unscrambler Menu Options The Unscrambler Program Operation Move the cursor around and let it rest on the part of The Unscrambler you want help about Left clicking on an object will give you specific help about that object as detailed in the table below When Help button is activated click on to get help on Area you click on Gives you Help about Menu option The corresponding command Toolbar button The corresponding command Plot in the Viewer Plot interpretation similar to lt F1 gt Other parts of the program The corresponding feature e g The Workspace or the table of contents Help Tutorial Exercises Starts the Tutorials that guide you through several application examples How to Use the Tutorials Each tutorial starts with a presentation of the application example Read carefully so as to understand the context of the application and the nature of the data The next chapters of the tutorial are devoted to practical tasks The
142. To Sort Select Variable Sets To Sort Dialog This dialog has the same shape for sorting Samples or Variables by sets Dialog Select Sample Sets To Sort Select Sample Sets To Sort Select the sets to sort from the list of available sets or define new sets using the Define Button Selected Sets Available Sets Name Interval Name Interval Training Set 1 103 Raw data 1 106 Prediction 107 118 Define The Unscrambler Program Operation The Modify Menu e 235 Sort Samples by Sets Select the sample set that must be placed in the first rows of your data table from the list of Available Sets and click on the Add button The chosen set is moved to the list of Selected Sets Proceed similarly to select the second sample set which will be placed in the following rows of the table then the third sample set etc Note After sorting samples that do not belong to any selected set will be displayed at the end of the table Sort Variables by Sets Select the variable set that must be placed in the first columns of your data table from the list of Available Sets and click on the Add button The chosen set is moved to the list of Selected Sets Proceed similarly to select the second variable set which will be placed in the following columns of the table then the third variable set etc Note After sorting variables that do not belong to any selected set will be displayed at the end of the table Common Features Overlapping set
143. Unscramble README TXT 25 When the setup is complete click Finish 18 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS InstallShield Wizard InstallShield Wizard Complete Setup has finished installing The Unscrambler on your computer 26 Start The Unscrambler Client from the Start menu A login dialog will appear Type in your user name and password to start the application Troubleshooting under Installation Symptom 1 InstallShield Error Message When running the setup exe file of The Unscrambler or of The Unscrambler License Server installer the following message appears The Unscrambler Program Operation Troubleshooting under Installation e 19 InstallShield Return code 2147024894 E 1608 Unable to create InstallDriver instance 3 To solve this problem double click on file IDriver exe located in path C Program Files Common Files InstallShield Driver 8 Intel 32 or similar location Intel 32 File Edit View Favorites Tools Help Q sxx Q Si JO Search 1 totes EI KE x i Ek Address C Program Files Common Files InstallShield Driver Intel 32 v Go Folders X Name Size Type Date Modified CA _I5RES1033 dll 284KB Application Extension 05 03 2003 18 45 EM ipriver2 exe 632 KB Application 28 03 2003 18 21 632KB Application 28 03 2003 18 21 3 1Scripts dll 232 KB Application Extension 01
144. Unscrambler requires that one person is appointed as supervisor system manager The supervisor s main task is to maintain the user accounts In a single user installation of The Unscrambler you are your own supervisor The supervisor must log in after installation and define the users who are allowed access to Unscrambler before they can begin to work with the program In a single user installation this simply means that you have to log in as Supervisor in order to modify the default user Guest and give it a name of your choice usually your own name 22 e Installation and Configuration The Unscrambler Program Operation dp The Unscrambler User Manual Camo Software AS System Configuration in Practice Start The Unscrambler and log in as Supervisor by right clicking on the caption bar in the Startup dialog or pressing lt Ctrl gt lt Shift lt S gt see the figure below Dialog The Unscrambler Startup The Unscrambler Startup ChieShifteS trl S hift S Users Move Close Alt F4 Help This triggers the Login dialog where you may log in as Supervisor by typing in the correct password The default supervisor password at delivery is SYSOP User accounts are maintained from File System Setup The System Setup dialog consists of 3 sheets e Users e Import Export e Directories System Setup Dialog Users Sheet As supervisor you can add a new user delete an existing user or modify user properties
145. Variable Statistics 3 D This option is disabled for all 3 D data tables no matter which layout they have 152 e The Unscrambler Menu Options The Unscrambler Program Operation Useful tips 1 To display Variable Statistics for an OV table select the whole table then drag and drop it unto an empty spot of the workspace alternatively use File Duplicate As 2 D Data Table You may then display Variable Statistics on the resulting 2 D table 2 To display Variable Statistics for an ON table two solutions are possible Duplicating the table as 2 D and displaying Variable Statistics will give you a summary where Primary and Secondary samples are taken into account together If you are interested in Variable Statistics across individual Primary resp Secondary samples here is what you can do Select the variable you want to focus on Choose File Convert Vector to Data Table The resulting 2 D table has as many rows as you had Primary samples and as many columns as you had Secondary samples Display Variable Statistics on that table if you want a summary of each Secondary sample across all Primary samples Display Sample statistics if you want a summary of each Primary sample across all Secondary samples Sample Statistics Dialog and Variable Statistics Dialog The dialogs Sample Statistics and Variable Statistics let you compute simple statistics for all samples or variables in your data table Dialog Sample Statistics
146. ables are included in the model in this case it is necessary for you to go through a screening stage to identify the influential variables before you start an optimization Define Design Purpose Next Dialog Once you have chosen the purpose of your design you may proceed by clicking Next You will then access the next dialog Design Type Mixture Design Wizard Design Type Mixture This dialog is specific to mixture designs with a simplex experimental region In this dialog you can choose the kind of Mixture Design that you want to use Simplex Lattice Centroid or Axial as well as the settings for these designs Lattice degree inclusion or not of Interior or End points A recommended choice is displayed and unless you have a particular reason not to follow the advice you should follow it 112 e The Unscrambler Menu Options The Unscrambler Program Operation Design Wizard Design Type Mixture The mixture design recommended to achieve your purpose is axial design without end points Unless you are familiar with mixture designs and have a good reason for changing this choice we recommend that you keep this default The number of mixture points for the chosen design is listed hereafter If your design includes process variables some of these paints will be combined with the process part D optimally Number of Mixture Points fe E Reset Type of Design C Simplex Lattice Design Degree PE E glesi
147. ables of both types make sure to mark all the variables X and Y you want to keep active in your model Task Extract Data from Marked Extracts data from the marked objects in the plot and displays them in an Editor window Task Extract Data from Unmarked Extracts data from all objects other than the ones you have marked and displays them in an Editor window 288 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS The Results Menu Results All This command launches the Results dialog which gives you an overview of all your data and results files The dialog is configured to display Unscrambler files of all types Results Dialog The Results dialog gives you an overview of your data and or results files together with some useful buttons Dialog Results Lookin JC USE Test data el Name Type Creato Modified la Jam micro 41D Unse PLS1 DG 02 04 2006 16 54 gi PLS1 spectra 25x2 Unse PLS1 DG 02 04 2006 10 14 File name H am micro 41D View Files of type Regression M Close Tl Mine only Export Information JAM MICRO a Print irectory C MY DOCUMENTS THE UNSC ALBEE el Creator Dominique Guyot Warnings ate 02 04 2006 16 54 Software Version v9 5 0 5 Calibration Method PLS1 alidation Method Cross Validation Model Size Mi G ode ize cro x Extracted C Notes Variance
148. active plot contains marked samples If you wish to select variables make sure that the active plot contains marked variables Note2 In regression results when using this option from any active variable plot the following rules apply e If only some X variables are marked the Recalculate With Marked option runs a new model including only the marked X variables and all Y variables e If only some Y variables are marked the Recalculate With Marked option runs a new model including only the marked Y variables and all X variables e If both some X variables and some Y variables are marked the Recalculate With Marked option runs a new model including all marked variables and excluding all unmarked variables The Unscrambler Program Operation The Task Menu e 287 Task Recalculate without Marked Makes a model on all objects samples or variables other than the ones you have marked in the plot The model parameters are the same as in the model used to make the result plot where you marked these objects That means for instance that if the plot shows the results of a PCA a new PCA will be run excluding all marked objects In the PCA dialog the list of marked objects is automatically displayed in the Keep Out of Calculation field The command can be used for marked samples for marked X variables and or for marked Y variables Note The recalculate function applies to the samples or variables marked in the currently active plot
149. aeesseeereees 41 Importing Data into The Unscrambler Using Drag n Drop 41 Transferring Data from Spreadsheets into The Unscrambler 0 cite eee ceeeeseeeeereeeees 41 Transferring Unscrambler plots to Another Application 0 cic cseeeeeseeecseeeceseeeeneeseens 42 Import Export in Farbe 5 2 o65 tacs eene isesecd petegenae ara eA e Wate E ERSAT 42 How To Import Results or Data O Di 43 How To Import Results or Data OD 43 How TO Export Dataiscsssescsscavscivstssshevesdasssenedet annin a EErEE EEN gege 43 How To Export MGdels ccccccscossseveccsceesesvecsesssosecosisescesstevenssiesesstectisnssessenstevensesess ieevsedss ES 43 The Unscrambler Menu Options 45 Th Fil MGM zeegt eet eege ege eaiists cvs seses E E deeg NEEN 45 Bile EE 45 Piles E 48 Filer EH 48 Pile geet egenen Deeg tseecuech esieenl 49 SUE 49 SUE 49 SU dE 50 File Import Unscrambler Results 51 File Import Unscrambler Data 52 File ae e E US E UC 53 Pile Import US RE 53 File Import Unscrambler ASCI csccsccsssssscssscssesccsscssscnsessscssessesescnsovssesseesseesereues 53 File Imp ort AS LE 53 Pil lmniport EL EE 55 File pe Excel i ss cscssscschevssvncseheesenavedenecescerspnsspeassoncessVansdeueseeoanadssasbepnssbedbanesiestecnenneseles 55 File Import Database koeren eresie E Era EENES EIEEE K EEES 56 File Impott JCAMP DX EE 59 File Import KEE 61 File Import INSAS EE 61 Pile ee E GE 61 File Imp or
150. age of not generating any missing value at the ends of the spectrum Dialog Norris Gap Derivatives Norris Gap Derivatives x m Scope OK Samples Cancel dl Samples Define Help di Variables Spectra 100 Define Parameters Derivative order 2nd Derivative X Gap size number of points 1 2 In the Scope field select the samples and variables to be transformed If necessary press Define to enter the Set Editor and define new sets In the Parameters field choose the Derivative order i e whether you want the Ist 2nd 3rd or the 4th derivative of the samples from the drop down list Then select the required Gap size width of the interval between the two values used for differentiation 228 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Modify Transform Derivatives Gap Segment Camo Software AS This method computes derivatives of up to the 4 order It is a more complex version of the Norris gap method that includes an additional segment option for smoothing Dialog Gap Segment Derivatives Gap Segment Derivatives x Variables Spectra 100 D Define Define m Parameters Gap size number of paints Segment size number of points Derivative order 2nd Derivative HS fi OK ai Cancel SS Help In the Scope field select the samp
151. aking the absolute value gives equal meaning to positive and negative correlations due to which anti correlated samples will get clustered together Un centered Correlation distance This is the same as the Pearson correlation except that the sample means are set to zero in the expression 256 e The Unscrambler Menu Options The Unscrambler Program Operation for un centered correlation The un centered correlation coefficient lies between 1 and 1 hence the distance lies between 0 and 2 e Absolute Un centered Correlation distance This is the same as the Absolute Pearson correlation except that the sample means are set to zero in the expression for un centered correlation The un centered correlation coefficient lies between 0 and 1 hence the distance lies between 0 and 1 e Kendall s tau distance This non parametric distance measurement is more useful in identifying samples with a huge deviation in a given data set Note If the data set contains any sample with a standard deviation of zero i e no variation among its variables the 4 correlation methods should be avoided Once you are satisfied with your choices hit OK The clustering results are displayed in the data table as a Category variable called lt Distance type gt _SOD lt value of sum of distances gt containing the ID 1 2 N of the cluster each sample has been affected to For instance if the clustering was performed using the Euclidean di
152. al Camo Software AS The View Menu View Sample Statistics Displays a few descriptive statistics about your samples in a new Editor You first access the Sample Statistics dialog where you must select a Variable set The computed statistics are e Number of missing values e Minimum e Maximum e Mean Standard deviation e Skewness The results are displayed in a new Editor This new Editor behaves as a slave of the source data Editor As a consequence rows marked in the original data table will also be marked in the Sample Statistics table and vice versa An example of Sample Statistics results oe a E EE EE EE Se Calibration a 0 0000 20000 10 0000 7 3333 46188 0 3849 Calibration 2 0 0000 40000 20 0000 14 6667 9 2376 0 3849 Calibration 3 0 0000 10 0000 32 0000 19 3333 11 3725 0 2679 Calibration 4 opp 20 0000 440000 30 6667 12 2202 0 2078 Calibration 5 0 0000 0 0000 0 0000 0 0000 m Calibration ie 0 0000 3 0000 15 0000 11 0000 69282 0 3849 Calibration 7 0 0000 50000 31 0000 15 6667 13 6137 0 3025 Unknown fe Unknown Je Unknown fa 1 0000 12 0000 21 6000 16 8000 6 7882 2 0722e 07 1 0000 20 0000 424000 31 2000 15 8392 7 2497e 09 1 0000 13 2000 18 0000 15 6000 3 3941 2 0722e 07 The usual File Edit View Modify and Task operations are not allowed on this slave Editor only Plot is possible Besides closing the Editor contai
153. alog Variance Dialog This dialog is accessed by clicking Variance from the Classification dialog the Prediction dialog or the Get Model dialog Dialog Variance Variance Model Name D OK Suggested Number of PCs ba Cancel Help Calibration Validation PC x x 0 36 0 39 2 78e 02 3 42e 02 3 95e 03 5 61e 03 1 21e 03 3 16e 03 3 17e 04 2 22e 03 1 31e 04 1 86e 03 The relevant variances X Y Cal Val are displayed for each PC Click the button if you want to switch from absolute values to percentages Click OK or Cancel to go back to the main dialog Task Predict This task allows you to predict response values from a regression equation Depending on the nature of the data table displayed in the Editor upon choosing Task Predict the following dialog is accessed e For a regular 2 D data table the Prediction dialog shown hereafter allows you to use a regression model of type PLS1 PLS2 or PCR 282 e The Unscrambler Menu Options The Unscrambler Program Operation e For a3 D data table the Prediction Three Way PLS dialog allows you to use a tri PLS nPLS regression model This dialog is described in section Prediction Three Way PLS Dialog Note What follows applies to 2 D data only The Prediction dialog contains four sheets Samples X variables Y reference Pretreat Vars as well as some general options at the bottom Dialog Prediction Prediction Samples X variables Y teferen
154. along the X Y and Z axes Then select which variable set or sample set to take into account For instance in the dialog example shown in the above figure the same data file Calibdata has been chosen as source for X Y and Z The vectors to be plotted against each other are variables number 5 7 and 9 The samples which will appear as points on the plot are taken from sample set Training Source File Is A Result File In this case you must specify from which matrix stored in the result file you wish to plot your X vector Y vector and Z vector Then select which dimension of the matrix you wish to pick elements from the dimension which you have not selected is the one that will vary in the plot Note You may select more than one vector at a time as X Y or Z or all each triplet of vectors will appear as a separate series of points on the plot Once the plot is displayed in the General Viewer you may add more series of points with the command Edit Add Plot This opens a new dialog where you can select more vectors to be plotted together with the ones already displayed Plot Histogram General View This command selected from the General Viewer launches a dialog where you can specify a histogram plot of any vector of your choice from any Unscrambler file Select the file from which you want to plot in the Source field If you do not remember the file name click the Browse button to launch the Results dialog
155. ambler Menu Options The Unscrambler Program Operation Scores and Loading Weights Dialog Two Plots Sheet This sheet lets you specify two plots a Score plot in sub view 2 and a Loading Weights plot in sub view 3 e Choose the PCs for which you want to look at the plots in the Components field e Inthe Samples field select whether to plot Calibration and or Validation samples e Inthe Loadings Mode field select whether to see Primary or Secondary X variables e In the Loadings field select whether to see X or Y loadings or both Scores and Loading Weights Dialog Four Plots Sheet On this sheet you can specify four plots two Score plots plot 1 and plot 2 in sub views 4 and 5 and two Loading Weights plots plot 3 and plot 4 in sub views 6 and 7 e Specify four PCs in the Components field the two axes common to plots 1 and 3 first Scores plot and first Loading Weights plot and the two axes common to plots 2 and 4 e Inthe Samples field select whether to plot Calibration and or Validation samples e Inthe Loadings Mode field select whether to see Primary or Secondary X variables e In the Loadings field select whether to see X or Y loadings or both Scores and Loading Weights Dialog Bi plots Sheet Here you can specify one plot which displays scores and X loading weights projected onto the same set of 2 PCs e Choose the PCs for which you want to look at the plots in the Components field e Inthe Loadings Mode field sel
156. ame string indicates the start of a character sequence The next click indicates the end of the name string All characters in between are selected If you want to select characters that are not in sequence hold down the lt Ctrl gt key when you select the characters you want in the label with the mouse Deselect one by using lt Ctrl gt and clicking again Edit Insert Draw Item You can add a drawing to the plot using this option It is possible to draw with two different shapes N e Line e Text T Change Item Properties once it is Drawn Mark the item by pressing the lt Alt gt key and clicking with the left mouse button at the same time You can then delete the drawing item or change its properties The item can also be moved around the plot by dragging it while the lt Alt gt key is pressed down Once the item is selected use Edit Item Properties to change its properties style font color A marked draw item can also be moved by using the arrow keys The Unscrambler Program Operation The Edit Menu e 143 Edit Insert Draw Item Line N Allows you to draw a line on a plot Click at the point in the plot where you want the line to start and hold the left mouse button down as you draw the line The properties of the line can be changed at a later stage by marking the item Press the lt Alt gt key and click on the line and selecting Edit Item Properties For details on the Line Properties dialog looku
157. amely the ANOVA table for the effects the normal probability plot of the residuals a contour plot and a landscape plot of the surface itself This overview is displayed for one Y variable For the contour and landscape plots the most significant X variables the ones with the largest absolute B coefficients are used if the number of X variables is larger than two Axes not shown are set to their average values You only have to make one choice in the Response Surface Overview dialog Choose which Y variable to study in the Y variable field Plot Analysis of Variance This plot displays the ANOVA table which tells you about the significant effects in the model If you are displaying results from an Analysis of Effects with variables having more than two levels the same plot dialog also enables you to plot multiple comparisons This is a comparison of the average responses for the different levels of a design variable which is useful in determining which levels of a variable are responsible for a possible significant effect The ANOVA table shows degrees of freedom sums of squares mean squares f values and p values for all variables as well as interactions if they are used in the model The Multiple Correlation and the r square are also presented If the results come from a response surface a model check and a lack of fit test will be shown below the ANOVA table for the effects In the latter case intercept terms and max min saddle points a
158. ancel Help File name X vars FirstX var Last Zwar Step Y vars 1101 191130184001 1101 4nalyzer 5036 Channel 1101 Analyzer 5036 Channel 1101 Add file name to sample name Import Y Variables Iw Import Predicted Y Variables File information System Name System 1 AnalpzerlD 4037 Channel Number 1 Analyzer Temperature 34 7461 Process Steam Name Channel 1 The CLASS PA amp SpectrOn Import dialog box displays a list of files from which you may import Guided Wave data File name number of X variables names of the First and Last X variables Step and number of Y variables are displayed for each file If necessary click the Browse button close to the Look in field in order to access files from a different folder Once you have selected a file from the list information about the file contents is displayed in the File information field at the bottom helping you check that you have chosen the correct file Multiple selections are possible using Shift Click or Ctrl Click You may select 3 additional options e Add file name to sample name e Import Y Variables e Import Predicted Y Variables Click OK to validate your choices and perform the import The Unscrambler Program Operation The File Menu e 63 The Unscrambler User Manual File Import Indico This option lets you import data from files created with the Analytical Spectral Device software Indico Source files with the following forma
159. and RMSEP which contains one sheet for each of the possible plots listed above Variances and RMSEP Dialog X or Y variance Sheet This dialog sheet enables you to select either X variances or Y variances and display them as a single plot Variance is plotted against component number You can choose between calibration and validation samples or both Explained variance is selected by default you can later change that by using View Source In the field Variables you select which variables you want to display in the plot You can choose between variance for the X variables or the Y variables To select All variables click this button and to enter an Editor where you can select variables click Select Tick the Total box to display the total variance at the same time You can decide whether you want the variance to be plotted for calibration or validation samples respectively or both by ticking the appropriate box es in the Samples field Calibration variance measures the fit of the model to the calibration data The validation variance is an estimate of the fit of the model to new data not present in the calibration and is based on the validation Plot both measures to obtain full information about model fit The dialog displays a preview screen Double click on this to change the size of the plot 178 e The Unscrambler Menu Options The Unscrambler Program Operation Variances and RMSEP Dialog X or Y variance Sheet Three Way PLS R
160. arate window as shown below The Unscrambler Program Operation The View Menu e 171 The Unsc The MCR Message List window MCR Message List E xj Type Previous Setting 1 100 No recommendation 2 100 No recommendation H 100 Change sensitivity to pure components is recommended 4 No recommendation es Check for any system recommendations e Recommendations of type 1 3 have to do with the value of field Sensitivity to Pure Components in the Multivariate Curve Resolution MCR dialog Column Previous Setting displays the value you chose when specifying the MCR e Recommendations of type 4 are more general e g Data normalization is recommended View Outlier List This toggle selection lets you show or hide the outlier list This dialog displays the outlier warnings that The Unscrambler generated during analysis of the data tables or results you have plotted in the active Viewer The tests that are performed are described in the Method References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices The dialog consists of 7 columns Below is a brief description of each column in the Outlier List dialog The columns in the Outlier List dialo Columns Description he PC number for the current test is shown The tests are calculated separately for each PC in the model he sample number is given for all positive tests that are relat
161. ariable amp Sample Use this option to exchange Secondary variables and Samples in a 3 D table with layout OV As a result e Former Secondary variables are now Samples e Former Samples are now Secondary variables e Variable names and Sample names are updated accordingly e Former Primary variables are still Primary variables The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z Modify Swap Samples amp Variables Pri Sample amp Sec Sample Use this option to exchange Primary samples and Secondary samples in a 3 D table with layout ON As a result e Former Primary samples are now Secondary samples e Former Secondary samples are now Primary samples e Sample names are updated accordingly e Former Variables are still Variables The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z Modify Swap Samples amp Variables Pri Sample amp Variable Use this option to exchange Primary samples and Variables in a 3 D table with layout OV As a result e Former Primary samples are now Variables e Former Variables are now Primary samples e Variable names and Sample names are updated accordingly e Former Secondary samples are still Secondary samples The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z Modify Swap Samples amp Variables Sec Sample amp Variable Use this option to exchange Secondary samples and Variables in a 3 D table wi
162. ase sample selections on the levels in some variables The variable is defined in Variable and the level in Level You determine whether the selected samples should be the samples with or without the specified level s Selecting Samples lets you specify the samples you want to select add or remove Adjacent samples are indicated by a hyphen e g 4 8 means samples 4 5 6 7 and 8 while commas are used as delimiters in multiple selections e g 1 4 8 10 15 25 If you want to select all samples click the All button to the right of the field Edit Select Variables In the dialog Select Variables you may select the variables you want to work with as the variable set Currently Selected Variables in your analysis You can increase or reduce the number of variables you are working with at the moment The field Selection method is identical to the one in the dialog Select Samples In the Select using field you select to use either an existing set as in the Select Samples dialog or variables which you specify Use the button All to specify all variables Adjacent variables are specified with a hyphen eg 4 8 commas are used to separate multiple intervals eg 1 4 8 10 15 25 Edit Select All You can use this command to select the entire matrix The next operation you choose will be performed on all elements in the matrix Note The Unscrambler always works with either rows or columns This also applies when the who
163. ased on the information found in the NSAS data file reflecting any transformations applied to the data before the NSAS data file was created For this reason you should not change any default values unless you are very confident in what you are doing Any transformation applied to the raw data using The Unscrambler must be manually specified For more information about the correspondence between Unscrambler and NSAS transformations see the Chapter NSAS Transformations Press the Log button on this sheet to enter a Log where you can make notes about your export File Export Model Tracker Tracker models are written in a format used by Foss instruments You can write models in both CAL and PIF format Specify the folder in which you want to save the model in the Folder field in the Export Tracker Model dialog Use the Browse button if you do not remember the folder name Then in the Model name field enter the name you want the model file to have on the disk In the PCs field enter the number of PCs you want the model to be made up of in the exported model file This will usually be the optimal number of components you have found during your work in The Unscrambler There are two types of Tracker models CAL and PIF The Unscrambler Program Operation The File Menu e 79 Camo Software AS The Unscrambler User Manual File Convert Vector to Data Table The menu option File Convert Vector to Data Table allows you to conv
164. at will refer to the set in other dialogs when you want to make models plot data etc It should therefore be evocative of the nature of the variables in the set Variable Sets may be of type Spectra or Non spectra Results from models where sets marked as Spectra are used have different default plots For instance the loadings are plotted as line plots instead of 2D Scatter plots Make your choice between Spectra or Non spectra in the Data type field 246 e The Unscrambler Menu Options The Unscrambler Program Operation he The Unscrambler User Manual Camo Software AS Dialog New Variable Set New Variable Set sl rM General Name fouaity Data type Non Spectra z Cancel Help ike m Interval valid range is 1 through 19 fe 7 1 0 Select Special intervals C All variables Selected Variables inthe eunenty active Editar Every E variables counting from fi a Update m Interaction and square effects In the Interval section enter the variable range Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10 13 27 Pressing the Select button next to this field launches an Editor the Select Variables dialog which allows you to select variables by clicking on them In the Special intervals field you can enter the set interval by using the currently selected variables in the Editor Every nth variable can also be entere
165. ate the marker for example This requires a Indication Name designed data table Position on file The basis for the numbering is the raw data Value of Variable Use the value of a specific variable as marker for the samples Category variables are useful with this option Use the tick boxes to select the part of the level names to be displayed in the plot If you take the options Name or Value of Variable use the tick boxes to select which part of the name level name or numerical value shall be displayed on the plot 136 e The Unscrambler Menu Options The Unscrambler Program Operation The first mouse click on a character in the name string indicates the start of a character sequence The next click indicates the end of the name string All characters in between are selected If you want to select characters that are not in sequence hold down the lt Ctrl gt key when you select the characters you want in the label with the mouse Deselect one by using lt Ctrl gt and clicking again If you take the option Value of Variable you need to specify which type of variable shall be used as a marker a Levelled variable an X variable or a Y variable from the drop down list Then you may either type in the variable number in the open text field or click Select to access the Select Variable dialog This dialog is like a small editor showing you the available variables among the chosen variable type and letting you select one of thes
166. ath Vision Export Dialog General and Instrument Parameters Sheets These sheets are identical to the sheets in the NSAS Export dialog see File Export Model NSAS Vision Export Dialog Math Sheet This page of the property sheet is used to specify which transformations have been applied to the data before calibration has been performed This is important for use of the exported model for predictions as spectra used as input to the prediction must undergo exactly the same transformations as the calibration spectra Vision allows for up to three transformations specified in the fields 1st NSAS transformation 2nd NSAS transformation and 3rd NSAS transformation respectively The first transformation is used for the sole purpose of indicating whether the spectra has been corrected for reference changes within the instrument or not The second and third transformations may be any Vision math type including None Select from the drop down list For some of the math types one or more of the parameters Segment size Gap size or Divisor point will be enabled These will be written to the model file exactly as shown Use the spin buttons to specify them Default values for the transformations are taken from the Instrument Information block in The Unscrambler s result file which is again copied from The Unscrambler s data file If the data file was originally imported from an NSAS data file the Instrument Information block was created b
167. ations Accessed with the command Plot Estimated Concentrations this command produces a line plot of the Estimated Concentrations in your Multivariate Curve Resolution MCR results The Estimated Concentrations dialog contains only one sheet called General Dialog Estimated Concentrations Estimated Concentration E x General Pon Ch Cancel Help m Number of Components i Suggested no of Components is 4 p o S Concentration Profiles fi 2 Al In the Number of Components field you may tune up or down the number of components for which the Estimated Concentrations will be plotted In the Concentration field you choose which profiles to display These are the Estimated Concentration profiles of each constituent plotted across all the samples included in the analysis Use the All button to display the profiles of all the estimated constituents Plot Estimated Spectra Accessed with the command Plot Estimated Spectra this command produces a line plot of the Estimated Spectra in your Multivariate Curve Resolution MCR results The Estimated Spectra dialog contains only one sheet called General In the Number of Components field you may tune up or down the number of components for which the Estimated Spectra will be plotted In the Spectra field you choose which profiles to display These are the Estimated Spectra of each pure constituent plotted across all the variables include
168. atural and intuitive way of computing a distance between two samples It takes into account the difference between two samples directly based on the magnitude of changes in the sample levels This distance type is usually used for data sets that are suitably normalized or without any special distribution problem Manhattan distance Also known as city block distance this distance measurement is especially relevant for discrete data sets While the Euclidean distance corresponds to the length of the shortest path between two samples i e as the crow flies the Manhattan distance refers to the sum of distances along each dimension i e walking round the block Pearson Correlation distance This distance is based on the Pearson correlation coefficient that is calculated from the sample values and their standard deviations The correlation coefficient takes values from 1 large negative correlation to 1 large positive correlation Effectively the Pearson distance lies between 0 when correlation coefficient is 1 i e the two samples are most similar and 2 when correlation coefficient is 1 Note that the data are centered by subtracting the mean and scaled by dividing by the standard deviation Absolute Pearson Correlation distance In this distance the absolute value of the Pearson correlation coefficient is used hence the corresponding distance lies between 0 and 1 just like the correlation coefficient T
169. automatically written to the Log sheet The Unscrambler Program Operation The File Menu e 85 Camo Software AS Dialog Properties Notes Sheet DESIGN_CHEESE__RESPONSES Properties E General Notes Loa Instrument Constraints CHEESE Sample Center a was made with 42 of Dry Matter instead of the planned 38 Therefore this sample will be kept out from Statistics and Analysis of Effects It may be studied in a PCA later on M Word wrap The Unscrambler User Manual Properties Dialog Log Sheet and Transformations Sheet This sheet is called Log for data files and transformations for result files Data Table A Log of all transformations and other changes to the data table is displayed in this sheet Dialog Properties Log Sheet Design Cheese Properties x General Notes Log Instrument Constraints Action User Time Parameters P Statistics P Statistics P Analysis of Effects Insert Variable s P Statistics D Analysis of Effects P PCA Cancel Apply Help VL 03 10 03 13 41 VL 03 10 03 13 41 VL 03 10 03 13 41 VL 03 10 03 13 42 VL 03 10 03 13 42 VL 03 10 03 13 43 VL 03 10 03 13 43 ID C SYSTEM ID C SYSTEM ID C ASYSTEM RIGHTNO 10 NL ID C SYSTEM ID C SYSTEM ID C SYSTEM 3 Empty 86 e The Unscrambler Menu Options The Unscrambler Program Operation he Unscrambler User Manual Camo Software AS The Empty bu
170. ay drag it with the mouse and move it around Drop it somewhere inside the workspace and it will float over the Viewer as a standalone window The Unscrambler with the Warning List displayed as a standalone window wv The Unscrambler Const PLS2 eles E File Edit view Plot Task Results Window Help icles alee EE Const PLS2 Variable v Total Const PLS2 Y var PC Qik Level 2 Milk Le For Help press F1 l4 Warnings Bi DG 296 e The Unscrambler Menu Options The Unscrambler Program Operation Shift to Other Window 1 2 This option lets you switch between all currently open windows A list at the bottom of the Window menu shows all open document windows identified with the name of the data or result file displayed in each window By clicking at one of the names the window in question will be activated The list is sorted in historical opening order first opened window at the top latest opened window at the bottom of the list The Help Menu Help The Unscrambler Help Topics This option opens the Help system on the Contents tab The Help window is divided into two halves The Main Help window displays a Help page The leftmost window contains the Table of Contents the Index and the Search engine Reading the Help System as a Book You can scan the Help system as a book by clicking your way through the chapters and sections Either access a chapter directly from the table of contents displayed to the left
171. c 2 Get Help on a specific menu option or button on the toolbar by activating the Help button La and clicking on the feature you wish to read about You may also open the help system by selecting Help The Unscrambler Help Topics or clicking on H this displays all the contents of the Help file From there you can click your way to the items you are interested in just as you would open a book Use the Index tab to search for keywords Several levels of help are available Click on underlined words to follow built in links to related help topics Read more about The Unscrambler Help System and its menu options in chapter p 297 Tooltips Whenever you let the cursor rest on one of The Unscrambler s buttons or icons a small yellow label pops up to tell you its function This is the quickest way to learn the functions of toolbar buttons 38 e General User Interface The Unscrambler Program Operation Communication with Other Software and Instruments Like most Unscrambler users you are probably already using many different programs in your daily work You might collect data using computers linked to your instruments or via data acquisition packages write your reports using word processors or use presentation programs to prepare lectures The Unscrambler can interact with these and many other applications in many ways The first chapter hereafter gives you an overview of the systems The Unscrambler is compatible with The nex
172. ce 163 projected 158 raw data 171 residual variance source 171 sample names 153 sample statistics 151 scaling 155 source 162 standard sample sequence 154 328 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual status bar 173 studentized residuals source 171 toolbar 173 trend lines 159 uncertainty test 160 validation source 163 variable statistics 152 viewpoint 156 X Y variable numbering 154 X1 variables source 170 X2 variables source 170 X variables source 170 Y residuals source 171 Y variables source 170 zoom in 156 zoom out 156 viewer 29 33 37 remember plot settings 92 Viewer remember plot settings 93 Viewer Toolbars dialog 173 viewpoint 156 Vision 79 visualize data 33 W warning limits leverage limit 266 ratio of calibrated to validated explained variance 266 ratio of calibrated to validated multiple correlation 266 ratio of validated to calibrated explained variance 266 ratio of validated to calibrated multiple correlation 266 sample outlier limit calibration 266 sample outlier limit validation 266 statistical condition number 266 total explained variance 266 variable outlier limit calibration 266 variable outlier limit validation 266 Warning Limits dialog 146 266 warning list 295 ways 314 weighting 314 1 SDev 311 passify 262 264 weights 259 260 262 passify 308 window cascade 294 close all 295 close current 294 copy to 292 ful
173. ce Pretreat Vars Variable Set Cancel x IW e Define Help Model Nene Era Number of Components 2 E Number of Pretreatments Iw Issue Wamings Waring Limits Variance Pretreat Prediction Dialog Samples Sheet Choose which samples to use in your calculation in the Sample Set field and enter any samples which you do not want to include in the analyses in the Keep Out of Calculation field For more details you may lookup the description of a similar dialog in Statistics Dialog Samples Sheet p 253 Prediction Dialog X variables Sheet First choose which Variable Set to use as X for prediction in the Variable Set field To define new variable sets click Define to go to the Set Editor If some of the X variables were kept out from the model at the calibration stage they will automatically be kept out at the prediction stage and their numbers will be displayed in the grayed out field Keep Out of Calculation This applies once you have selected the model to use see Prediction Dialog General Options Prediction Dialog Y reference Sheet Tick Include Y Reference on this sheet if you have reference measurements that you want to use in the prediction This will activate the Variable Set field where you specify which set to use as Y reference either by choosing an existing Variable Set from the drop down list or by defining a new in the Set Editor which you enter by clicking Define These reference valu
174. ch uses exactly one quarter of the window space 292 e The Unscrambler Menu Options The Unscrambler Program Operation Select which sub view you want to copy to by choosing a number The relationship between the numbers and the sub view layout is shown below Relationship between numbers and sub view layout Sub view Layout Sub view Layout BM 4 H 2 H 5 HH 3 RL s LR E EECH As a result of this operation the destination sub view becomes active i e current sub view the source sub view retains its plot but is no longer active Note Copy To erases the contents of the destination window Window Go To With this option you can go to a new sub view and select this as the current sub view i e activate it If you are currently displaying 4 sub views Go To 4 or 5 6 7 is strictly equivalent to clicking into sub view 4 or 5 6 7 Use Go To when you want to change your current sub view system for instance go from four sub views to two sub views Seven different sub views are available each of them may have some contents even though only a maximum of four are displayed at the same time The sub views are organized into three sub view systems 4 One sub view the Viewer window contains only one sub view number 1 which uses the whole window space 5 Two sub views the Viewer window is divided into two horizontal sub views numbers 2 and 3 each of which uses exactly
175. change the size of the plot Plot Regression Coefficients PCR and PLS The regression coefficients Bw scaled or B unscaled from a PCR or PLS regression analysis are plotted The scaled regression coefficients take into account the weighting options you have chosen when specifying your regression analysis They can help you identify the important X variables in relationship with a given Y variable In the Regression Coefficients dialog you have to specify the number of Components in addition to the other options Dialog Regression coefficients PLS Regression Coefficients Plot Type e Line C Matris Y variable fi MikLeve Components Suggested no is 5 Bal The dialog displays a preview screen Double click on this to change the size of the plot Note Once the plot of regression coefficients is displayed you may use options View Uncertainty Test Uncertainty Limits and Edit Mark Significant X Variables Only to study the significance of your X variables available only if you have chosen the Uncertainty Test option in the Regression dialog upon specifying your regression model The Unscrambler Program Operation The Plot Menu e 197 Plot Regression Coefficients Three Way PLS The regression coefficients Bw weighted scaled or B raw unscaled from a three way PLS model are plotted The scaled regression coefficients take into account the weighting options you have chosen when spe
176. cifying your three way analysis They can help you identify the important X variables in relationship with a given Y variable If you chose not to weight the X variables the Bw coefficients are identical to the Bs The Regression Coefficients dialog allows you to specify various options in addition to the type of coefficients Dialog Regression coefficients PLS Regression Coefficients Plot Weighted coefficients BW C Raw coefficients B Plot Type C Line variables e Primary x Ma Secondary X C Primary x Vs Vars C Secondary Ma Mars Y variable iD Severity Components Suggested no is 4 In the Plot Type field choose between a Line plot and a Matrix plot This will affect the options below In the X variables field select the variables to plot If your plot type is a Line plot choose between Primary and Secondary variables If your plot type is a Matrix plot you may choose to combine Primary and Secondary X variables for a fixed Y or to vary Y if your model includes several Y variables and combine it with either Primary or Secondary Xs In the Y variable field select the Y variable for which you want to display the regression coefficients In the Primary X variable field available for a Matrix plot of Secondary X Vs Y vars select the X1 plane for which you want to display the regression coefficients In the Secondary X variable field available for a Matrix plot of Primary X Vs Y vars
177. ck boxes INumber he basis for the numbering is the current Set Symbol e data points are represented by a symbol e g a small circle which does not give ny information about the point ariable Level he level of the sample is used to generate the marker for example This requires ndication Name designed data table Position on file he basis for the numbering is the raw data cl 138 e The Unscrambler Menu Options The Unscrambler Program Operation To the right of the radio button Name you choose which of the characters of the name you want to display The first mouse click on a character in the name string indicates the start of a character sequence The next click indicates the end of the name string All characters in between are selected If you want to select characters that are not in sequence hold down the lt Ctrl gt key when you select the characters you want in the label with the mouse Deselect one by using lt Ctrl gt and clicking again Options Dialog General Sheet for Histograms Accessed from Accessed from Edit Options for a histogram this sheet lets you define the plot layout Dialog Options General sheet for Histograms Options Ea General M Plot Header Options in the General sheet apply to all the data points in the plot The check boxes in the upper field are used to closer characterize the plot and are enabled and disabled according to which type
178. ckett Burman Central Composite or Box Behnken Define Design Variables e D optimal Non mixture Define Process Variables e Mixture Define Mixture Variables Select Existing Design Extend Next Dialog Once you have chosen the design you want to extend and clicked Next you enter a dialog called Extend a X design where X represents the type of design in question where all available extension options are listed Note1 Variables are shown in red if they are incompatible with the type of design extension that you have chosen If this happens select another extension or change the properties of the variable to fit the extension you want to do Note2 Box Behnken and Plackett Burman designs cannot be extended Note3 If any Design point is changed in a D Optimal or a Mixture design data table the design can no longer be extended This does not apply to changes in Reference samples Design Wizard Extend a Full Factorial Design This dialog gives you 3 choices regarding the extension of a full factorial design Delete Design Variables and Add More Levels More levels are typically added to a category variable when you find that some levels are missing Mark levels you want to keep out of the design and click the button between the two lists to remove them You enter the following dialogs Define Design Variables Define Non design Variables Design Details Randomization Details Last Checks 100 e The Unscrambler Menu O
179. column and a row is called a cell each cell holds a data value The rows and columns correspond to samples and variables respectively Samples and variables are identified by a number and a name 30 e General User Interface The Unscrambler Program Operation nscrambler User Manual Camo Software AS The Editor Active Cell and Cell Selection At any given time one cell in an Editor is active The active cell is marked with a frame Activate a cell by using the arrow keys to position the cursor or by clicking with the left mouse button on it You can also select a range of cells in the Editor i e one or more columns or one or more rows A whole row or column can be selected by clicking with the left mouse button on the sample or variable number the gray area between the names and the data table itself Keep the button down and drag the cursor to select more rows or columns Selecting a new range removes the last range To add new samples or variables to an existing selection and to make a range press the lt Ctrl gt key while you click on the appropriate samples or variables The range may be continuous or non continuous You can also deselect a sample or variable by pressing the lt Ctrl gt key while clicking on the object you want to remove from the range in toggle action This is only possible with the mouse Hold down the lt Shift gt key while you make the selection if you want to select a continuous block of samples or variable
180. crambler Program Operation The Unscrambler User Manual Camo Software AS File Duplicate As Non design The designed data table is copied to a normal data table in a new Editor All design information is lost in the process and the data is not locked anymore This means that you can modify the data with transformations and computations File Duplicate As Modified Design The current designed data table is used as a basis for creating a new design The Design Wizard is launched from the dialog Define Design Variables see p 104 with all settings of the current design It is then easy to go through the steps and make minor changes to the design instead of starting from scratch to make a new design that is similar File Duplicate As 2 D Data Table The menu option File Duplicate As 2 D Data Table allows you to convert a three way data table into a two way data table This feature is useful when you want to run an analysis meant for two way data e g PCA or PLS File Duplicate As 3 D Data Table The menu option File Duplicate As 3 D Data Table allows you to convert a two way data table into a three way data table This feature is useful when you already have a two way data table containing your unfolded three way data It creates a new 3 D table with the appropriate three way layout ONT or OV and reorganizes your data accordingly Dialog Duplicate as 3D Table x SES Samples All Samples 17 Define
181. cription can easily be applied to the Select Variables dialog by replacing sample with variable and row with column 250 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Select Samples Select Samples S B 1 6782 1 8111 1 6242 1 4680 S 1 8022 1 9797 1 7408 1 5463 2 1224 2 3275 2 0236 1 7649 2 8119 3 1621 26478 2 2136 3 0594 3 3984 2 8653 2 3977 1 4384 1 6088 1 4066 1 2436 1 7430 2 0329 1 6959 1 4271 1 70 1 6560 2 9554 3 6291 2 8051 2 1079 3 5590 41951 3 3567 2 5378 1 8226 21630 1 7665 1 4471 3 7 12 17 21 27 OK Cancel Help Click on the grey cell displaying Sample Number e g 12 to select one sample Include several Sample Numbers in one click if you want to perform a continuous selection Alternatively click on the first Sample Number you want e g 3 then scroll down press lt Shift gt and click on the last one e g 15 you want to include in your selection to achieve a continuous selection e g 3 15 Use lt Cirl gt click to select several samples in a non continuous range e g 12 17 Note In some cases multiple selection may be disabled Example selecting a Reference Spectrum as Background Information for EMSC the dialog does not allow for more than one sample in the selection You may scroll up down and right left whenever necessary At any time you may check your current selection by clicking Current the selected sample numbe
182. cs It may for example allow you to compare the distributions of variables of different scales within one plot These two transformations are column oriented the transformed values are computed as a function of the values in the same column of the table Dialog Center and Scale x Scope OK Samples Cancel Train 36 Define Variables Specific 147 Define Transformation V Mean Center Standard Deviation Scale The Unscrambler Program Operation The Modify Menu e 231 In the dialog Center and Scale select the Sample and Variable sets that define the matrix you want to correct in the Scope field The Define buttons take you to the Set Editor where you can define new sample and or variable sets In the Transformation frame two options are available e Mean Center within the selected sample and variable scope it subtracts the mean of all samples in the column i e variable mean from each cell of the column e Standard Deviation Scale within the selected sample and variable scope it divides each cell of the column by the standard deviation of the column i e variable Sdev Each option can be selected either separately or together with the other Note 1 If you only want to display the mean and standard deviation of the variables in your table use menu option View Variable Statistics Note 2 The Center and Scale transformations are not supported in a
183. cted A dialog will appear in which you select which Variable resp Sample Set to plot If necessary you may then click Define to enter the Set Editor where new sets can be defined Options in the Plot menu when one variable is selected in a 2 D data table Line 2D Scatter 3D Scatter Normal Probability Histogram Matrix Matris 3 D 32 e General User Interface The Unscrambler Program Operation ne Unscrambler User Manual Camo Software Plotting and Interpreting Results The Viewer In the Viewer data and results are visualized graphically in an interactive manner Whenever you make a plot it appears in a Viewer Every time the Viewer is mentioned throughout this manual and Help system we are referring to a window where a plot is displayed Several Viewers can be open at the same time In addition one Viewer can display several plots This is possible because the Viewer is divided into seven so called sub views organized as shown Organization of sub views Sub view Layout Sub view Layout The figure below shows a typical Viewer with sub views 4 7 The Unscrambler Viewer vw The Unscrambler RegrResult LL a x but File Edit View Plot Task Results Window Help l al x Disa sele Biel a del a m a wl as alal do Bipiot PC2 loading Weights and Y loadings Wl DCH T T T CH 10 05 D 05 RegrResult X expl 16 47 Y expl 6 12 T T T T T
184. ction by entering the model name in the Model Name field or by pressing the Find button next to it to enter the Get Model dialog The number of Primary and Secondary X variables in the prediction data set must be the same as the number of Primary and Secondary X variables used to build the model Specify how many PCs to use in the prediction with the spin button Press Variance to inspect the calibration and validation variances of the selected model This takes you to the Variance Dialog where the variances are shown for each PC in the selected model Toggle between residual and explained variance by pressing the button Task User defined Analysis This option has been developed at the request of a group of users who wanted to run their own analyses on 3 D data It works in a similar way to UDT s The option starts the User Defined Analysis UDA dialog where you can specify the scope of the analysis samples X variables of the primary and secondary type and Y variables which have to be stored into another data table where the 3 D structure is not used 286 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS The User Defined Analysis dialog User Defined Analysis i x Scope aK Samples mies Al Samples ST _ Define Cancel Help m General Variables Enter the primary and secondary variable data First variable mode 4 18 Second variable mo
185. d Click on the Obtain button to request your Unscrambler activation key on line If you have no internet connection contact your Camo office and provide us with your Unscrambler machine code Activating The Unscrambler 13 Type in your Unscrambler activation key into the relevant field of The Unscrambler Activation dialog then click the Activate button The Unscrambler Activation Unscrambler machine code Unscrambler 2 Solis shoe ben BSCNDBRBTDPRSFCKBCDMTBRBZTAKE tivate Currently The Unscrambler v9 5 is not activated c Click on the Obtain button to request your Unscrambler activation NGE key on line If you have no internet connection contact your Camo office and provide us with your Unscrambler machine code 14 The Unscrambler Startup dialog will appear The Unscrambler Startup Cancel Help Details The Unscrambler Program Operation Standalone Installation 1 User License Desktop Installation e 7 You may click the Details button to check the details of your license Details License Details The Unscrambler 9 5b Trial Yersion 29 Days remaining Click on buy for full version of The Unscrambler For any queries contact support camo com or Cancel support icamo no Defining User Name and Password User name and password are defined by logging in as supervisor in The Unscrambler Startup dialog Lookup chapter System Configuration in Practice for deta
186. d Select Existing Design 100 Design Wizard Extend a Full Factorial Desten eee eseeseeceeecseeseeeecaeeseeeeseeeeeneeas 100 Design Wizard Extend a Fractional Factorial Design eee ceeeeeeeseeee ceeeeeeeneceeteeeneenes 101 Design Wizard Extend a Central Composite Design 102 Design Wizard Extend a D optimal Non mixture Design 102 Design Wizard Extend a D optimal Mixture Design 0 0 0 ceeeseseeeeeseseesceeeseeeeceeeaaens 102 Design Wizard Extend a Lattice Design cee cee cseeeceeeeeseeeecneeseeeesseeneseeeaees caeeeeeaees 102 Design Wizard Extend a Centroid Design 103 Design Wizard Extend an Axial Design 103 Design Wizard Define Design Variables 0 ceccceceeeeseecesseseeseceeceeeaeeeesaeeaeeereneeeaeeees 104 Add Design Variable Variable Properties Dialog 104 Design Wizard Define Non design Variables 0 0 ccccecesesseeessseeceeeeeeseeneeeesaeeeeeeeaeenees 107 Design Wizard Define Mixture Variables 0 ccc ceeesessesesesceesceeeeeeeeeeesaeeseeessaeseneeeees 107 Design Wizard Define Process Vaables 109 Design Wizard Enter Multi Linear Constramnts ceeeessseeseeceeecseeseeeecseeaeeesseeseaeeas 109 Design Wizatd Define Model isver sesente ti cadecd geesde vai essees eair Eaa ET 111 Design Wizard Define Design Purpose ccesesceecceeeeeteeceeeeeeeeceeeaeeaceeeeeeeecesaeeeecneeaees 112 Design Wizard Design Type OMisture 112 Design Wizard Generate Design oo ieee eeesceeeceeeeeseceecee
187. d Plot 1 select which Primary X variables you want to display in the plot for a fixed Secondary X variable To select All Primary X variables click the All button Note that the Select button is disabled for three way data Tick the Total box to display the total variance at the same time Choose as SecX the fixed Secondary X variable that determines from which plane the selected Primary X variable s is are selected You may either pick up the Secondary variable s number or its name e The field Plot 2 works in a similar way and allows you to select which Secondary X variables you want to display in the plot for a fixed Primary X variable e You can decide whether you want the variance to be plotted for calibration or validation samples respectively or both by ticking the appropriate box es in the Samples field Calibration variance measures the fit of the model to the calibration data The validation variance is an estimate of the fit of the model to new data not present in the calibration and is based on the validation Plot both measures to obtain full information about model fit Variances and RMSEP Dialog RMSE Sheet This is a plot of the average prediction error for either the calibration or the validation samples In the plot you will find the Root Mean Square Error of Calibration RMSEC or Root Mean Square Error of Prediction RMSEP plotted against the number of components used in the model e If you select Calibration sample
188. d Wizard for The Unscrambler is launched InstallShield Wizard Welcome to the InstallShield Wizard for The Unscrambler The InstallShield Wizard will install The Unscrambler on your computer To continue click Next 21 Read the licence agreement carefully and click Yes to accept the terms of the agreement 16 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual InstallShield Wizard License Agreement Please read the following license agreement carefully Press the PAGE DO WN key to see the rest of the agreement IMPORTANT READ CAREFULLY The CAMO license agreement is a legal agreement between you the Licensee and CAMO PROCESS AS the Licenser and the company you bought the software from for the CAMO software product The software products covered by this agreement is The Unscrambler R The Unscrambler R Designer The Unscrambler R M A and The Unscrambler License Server hereafter referred to as the Program and includes computer software and associated media and printed materials and may include on line or electronic documentation By installing copying or otherwise using the Program you agree to be bounds by the terms in this license agreement If you do not agree to the terms in this license agreement promptly return the unused software to the place from which you E Do you accept all the terms of the preceding License Agreement If you choo
189. d in the analysis Use the All button to display the profiles of all the estimated constituents Plot Regression Overview Accessed with the command Plot Regression Overview this is a quadruple plot of some of the main results of a PCR or PLS regression namely the scores and loadings X and Y loadings residual Y variance The Unscrambler Program Operation The Plot Menu e 191 validation and predicted versus measured Y value as shown in the Plot field in the Regression Overview dialog Dialog Regression Overview Regression Overview Plot Scores Loadings OK Cancel Help Residual Y Validation Variance Predicted vs Measured Y variable 1l 4 PREFEREN Components Suggested no is 4 Scores and Loadings Vector 1 Vector 2 1 E 2 Predicted vs Measured 4 In the field Y variable you choose which Y variable to study and in the Components field you choose the PCs to be used as axes for Scores and Loadings and the number of components for Predicted versus Measured the latter should be the optimal number of PCs If the X variable Set used to generate the regression results is of type Spectra the Regression Overview dialog has a slightly different composition the 2D scatter plot of X and Y loadings is replaced with a line plot of the regression coefficients Regression Overview Dialog Three Way PLS Results Accessed with the command Plot Regression Overview when viewing thre
190. d into the set Specify the value of n with the spin buttons To enter the special set intervals into the Set Interval box press Update Finally you have the possibility to include Interaction and square effects in your new set in this field Press Select to launch the Interaction and Square Effects Selection dialog Note You can only include interaction and square effects in sets which include design variables New Primary Variable Set Dialog Allows you to specify a new Primary Variable set or modify its properties The Unscrambler Program Operation The Modify Menu e 247 Dialog New Primary Variable Set New Primary Variable Set General OK Name Reduced by 3 Cancel Datatype Spectra X Help Interval valid range is 1 through 665 1 4 7 10 13 16 19 22 25 28 31 34 37 40 4 Select Special intervals All variables e Every 3 j variables counting from 1 Update Interaction and square effects This dialog operates in the same way as the New Variable Set dialog see p 246 Note Interaction and square effects are disabled in Primary Variable sets New Secondary Variable Set Dialog Allows you to specify a new Secondary Variable set or modify its properties Dialog New Secondary Variable Set New Secondary Variable Set Pe General or Name All secondary Te Datatype Non Specta e Cancel Help Interval valid range is 1 through 31 fi
191. d is the one that will vary in the plot Note You may select more than one vector at a time as Abscissa or Ordinate or both each couple of vectors will appear as a separate series of points on the plot Once the plot is displayed in the General Viewer you may add more series of points with the command Edit Add Plot This opens a new dialog where you can select more vectors to be plotted together with the ones already displayed Plot 3D Scatter General View This option launches the 3D Scatter Plot dialog Here you must specify the file which you want to plot as X Y and Z in the X Y and Z axis fields If you do not remember the file name s click the Browse button in the field s in question to launch the Results dialog described in Results All There you can search for the right file Press Select to insert the correct file name in the 3D Scatter Plot dialog Dialog 3D Scatter Plot in General View 3D Scatter Plot OK X axis Calibdata Browse E Cancel Samples Variables Help Sam Set Training 28 1 28 E Y axis Calbdata Browse Samples Ze Variables Sam Set Training 28 1 28 EI Z axis Calibdata Browse E Samples Variables ER Sam Set Training 28 1 28 E The Unscrambler Program Operation The Plot Menu e 213 Source File Is A Data Table Once you have selected a source you must specify whether to plot a selected sample or variable
192. data is adjusted such that the data fill the region of the plot in all directions The results of a Range normalization on two different samples are listed below The Unscrambler Program Operation The Modify Menu e 223 Camo Software AS The Unscrambler User Manual Result of a Range normalization on two different samples 10 25 30 25 10 5 3 0 33 0 83 1 00 0 83 0 33 0 16 0 10 0 3 0 5 1 0 2 5 3 0 2 5 1 0 0 10 0 16 0 33 0 83 1 00 0 83 0 33 6 Peak normalization Peak normalization normalizes a sample as the ratio of each value by the value at a selected spectral point or wavelength The chosen spectral point usually the maximum peak of a band of the constant constituent or the isosbestic point is assumed to be concentration invariant in all samples Type in the number of the peak variable in field Var or use the Select button to select the peak variable directly from an Editor window Note If you peak normalize your calibration samples before building a model for later use in prediction or classification make sure that the same peak variable is selected when you normalize the prediction samples The equations used to normalize the samples are listed in the Method References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices Modify Transform Spectroscopic Transformation Sometimes you want to transform your spectra from one format to another Spectroscopic transformati
193. de fi 7 y variables Enter the name of the datasheet containing the y data Jun x y data fi 2 Note y data samples should contain the same no of rows as X data m Function NPls NPlsRun S Parameters Locate The analysis to be performed should be stored in a UDA file that you have previously installed on your computer See the User Defined Transformations section in the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices for a detailed description of how this works Use Locate to select your UDA from the list If it does not appear in the list of available UDT s and UDA s use Register to register it Task Recalculate with Marked Makes a model on only the objects samples or variables that you have marked in a given plot The model parameters are the same as in the model used to make the result plot where you marked these objects That means for instance that if the plot shows the results of a PCA a new PCA will be run on only the marked objects In the PCA dialog the list of unmarked objects is automatically displayed in the Keep Out of Calculation field The command can be used for marked samples for marked X variables and or for marked Y variables Note1 The recalculate function applies to the samples or variables marked in the currently active plot If you wish to select samples make sure that the
194. define new sets in the Set Editor If you want another resolution of your histogram you can change the number of bars to use in the plot using Edit Select Bars Use Edit Options to change the layout of the plot 176 e The Unscrambler Menu Options The Unscrambler Program Operation Plot Matrix E In this plot a two dimensional matrix is visualized The plot is useful if you want to get an overview of your data before you start your analyses as obvious errors in your data and outliers may be seen at once You can also want to take a look at this plot before you decide whether to scale your data or not Select the sample s and or variable s you want to plot in the Editor If you select samples you must specify which Variable Sets to plot in the Matrix Plot dialog and vice versa You can define new sets in the Set Editor by pressing the Define button If you have selected the entire data table in the Editor you can accept the default setup in the Matrix Plot dialog Four different layouts landscape bar map and contour are available Use Edit Options to change the layout of the plot Plot Matrix 3 D Choose Plot Matrix 3 D to display a matrix plot of one of the vectors in your 3 D data table With an OV layout you can produce a matrix plot of the Primary x Secondary variables for the plane represented by the selected sample With an O V layout you can produce a matrix plot of the Primary x Secondary samples fo
195. described in Results All There you can search for the right file Press Select to insert the correct file name in the dialog Source File Is A Data Table Once you have selected a source you must specify whether to plot samples or variables in the bottom right field Then select which sample s or variable s to plot either by entering it directly or by pressing Select to enter an Editor where you can mark your selection Ranges are indicated by a hyphen ranges or separate items are separated by a comma Source File Is A Result File In this case you must specify from which matrix stored in the result file you wish to plot Then select which dimension of the matrix you wish to pick elements from the dimension which you have not selected is the one that will vary in the plot For instance in the figure below the source file is a PCA result file called Blue The selected matrix is the scores matrix Tai and PC number 1 is selected which means that the plotted elements are the scores along PC for the 11 samples 214 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Histogram Plot in General View Histogram Plot Browse EI Cancel Matrix Tai 5 PCs 11 Samples Scores EI Help PCs Samples Source Blue Note Only one vector at a time can be plotted as histogram In the bottom right corner of the dialog a preview screen shows how the plot will look in the Gen
196. designed data tables When choosing this option you are asked for a number of variables to add After the insertion the total number of mixture variables in the table has to be three or more Note1 If you want to insert a mixture variable to a designed data table duplicate your table as non design by selecting the menu option File Duplicate As Non Design Then you will be able to insert mixture variables to your table Note2 You cannot insert Mixture variables into a 3 D table Edit Append Lets you extend your data table with new samples or variables You have 4 choices 1 Append Samples 2 Append Variables Options 1 and 2 are described in a common section Edit Append Samples or Variables 3 Append Category Variable 4 Append Mixture Variables Edit Append Samples See Edit Append Samples or Variables right below Edit Append Variables See Edit Append Samples or Variables right below Edit Append Samples or Variables You enter the dialog Append Samples or Append Variables where you use the spin button to select the number of samples or variables to add and click OK The Unscrambler Program Operation The Edit Menu e 125 The new samples are inserted below the current last sample The new variables are inserted to the right of the last variable and are of continuous non design type All the new data cells are filled with the value missing Append Samples or Variables 3 D
197. designs with a number of design samples equal to or close to your request from 1 to 2 experiments The best two designs lowest Condition No are displayed in the lower field Choose which one to use according to this rule the closer the Condition no is to 1 the better The condition number necessary to correctly explore your experimental region is automatically computed and displayed in the dialog text When reaching substantially higher condition numbers we recommend not using the design it will not cover the experimental region correctly but will describe some variables and let aside other ones Note If the two designs displayed do not suit you press Generate again The Unscrambler will generate four new designs and will compare them to the former ones before displaying the two best solutions Press Generate several times to give The Unscrambler a better chance to generate a good design for your purposes Generate Design Mixture When building a Mixture design this dialog is accessed in the two following cases e The experimental region is not a simplex 114 e The Unscrambler Menu Options The Unscrambler Program Operation 1scran mbler User Manual Camo Software AS e The experimental region is a simplex and the design includes Process variables These two cases are treated differently they are exposed hereafter The experimental region is not a simplex If the experimental region of your Mixture design is not a simp
198. dialog will be Design Details Note It is not possible to create a Centroid design if you have more than 6 mixture variables Design Wizard Extend a Centroid Design In this dialog you have the choice between the following 2 extensions Add Center and Reference Samples or Replicates Center samples and replicates are added to increase the reliability of the model Add reference samples to see how your design samples compare The next dialog will be Design Details Add Interior Points If they are not already included in your Centroid design you can choose to add Interior Points The next dialog will be Design Details Design Wizard Extend an Axial Design In this dialog you have the choice between the following 3 extensions Add Center and Reference Samples or Replicates Center samples and replicates are added to increase the reliability of the model Add reference samples to see how your design samples compare The next dialog will be Design Details Add End Points If they are not already included in your Axial design you can choose to add End Points The next dialog will be Design Details Extend to Centroid Design This choice first leads you through dialogs Define Non design Variables and Define Model Then you enter the dialog Design Type Mixture where the recommended extension design is Centroid Design with Interior Points Select this design and click Next The next dialog will be Design Detai
199. e e Any intermediate value for the star distance to center is also possible The design will not be rotatable Star Samples In optimization designs of the Central Composite family star samples are samples with mid values for all design variables except one for which the value is extreme They provide the necessary intermediate levels that will allow a quadratic model to be fitted to the data Star samples can be centers of cube faces or they can lie outside the cube at a given distance larger than 1 from the center of the cube see Star Points Distance To Center Test Samples Additional samples which are not used during the calibration stage but only to validate an already calibrated model The data for those samples consist of X values for PCA or of both X and Y values for regression The model is used to predict new values for those samples and the predicted values are then compared to the observed ones Test Set Validation Validation method based on the use of different data sets for calibration and validation During the calibration stage calibration samples are used Then the calibrated model is used on the test samples and the validation residual variance is computed from their prediction residuals Training Samples See Calibration Samples UDA See User Defined Analysis UDT See User Defined Transformation Uncertainty Limits Limits produced by Uncertainty Testing helping you assess the sig
200. e From the combo box choose between the three possible model sizes Full Compact and Minimum e The Full model which is the default contains all possible result matrices and thus can be opened in a Viewer and used in Prediction and or Classification on line or off line e The Compact model excludes the following result matrices from the model Eix X residuals Fiy residuals xWeighted and yWeighted weighted raw X and Y variables The Compact models have the same applicability as the Full models e The Minimum model only contains the result matrices needed for the Prediction and Classification analyses to work In addition three very small matrices they each only contain a vector of size A the number of principal components are included in order to maintain useful functionality in the file dialogs the variance plot the variance details dialog and the print details mechanism Preserving this functionality helps assessing the model quality when opening models for Prediction or Classification via file dialogs Note Restrictions on the Minimum Model The Minimum model does not contain enough result matrices to be opened in a Viewer When the calibration is finished the View button is hidden making the Close button the only allowed action in the Progress dialog When clicking this button you are asked to save the model as usual Later on if you try to open the model from the File Open or Results dialog a message will pop up in
201. e OBS and additional header information is stored in a header file HDR Both files must be present The variables Sets Filters and Constituents are defined Product Name and File Description is saved as instrument parameters in File Properties File Import GRAMS Here you may import files generated by GRAMS a software package used with some Galactic instruments You choose the appropriate file in the dialog GRAMS32 Import The Variable Sets Filters and Constituents are defined Product Name and File Description is saved as instrument parameters in File Properties File Import Matlab This option lets you import data from Matlab formatted files created in Matlab versions 5 x to 7 0 To Save Data for Importing Use the save command in Matlab e either save destinationfilename varl var2 e or save destinationfilename to save all variables in the workspace The Unscrambler Program Operation The File Menu e 61 This will create a Matlab formatted mat file For more help on using the save command type help save in Matlab To Import into The Unscrambler To import the file in The Unscrambler select File Import Matlab Select the destination filename in The Unscrambler to get the Import Matlab dialog box Select which selections represent the Data Sample names and Variable names The sample name and variable name variables must match the corresponding dimension of the data variable fo
202. e Several matrices can be imported together by combining either rows or columns Matrices that will not match the ones already chosen will be grayed out You can import several matrices either by rows or columns depending on whether your matrices contain the same number of variables row wise import or samples column wise import If the matrices contain sample and or variable names you can import them as well by selecting the appropriate option box File Import U5 Results Imports results files from earlier versions of The Unscrambler versions 3 0 5 5 If the file you want to import contains several matrices a dialog pops up to let you specify which matrices you want to import Note The Unscrambler recognizes the extensions UNM UNP and CLA Rename your files if they have other extensions File Import U5 Data Imports data files from earlier versions of The Unscrambler versions 3 0 5 5 If the file you want to import contains several matrices a dialog pops up to let you specify which matrices you want to import Note The Unscrambler recognizes the extensions UNS UNM UNP and CLA Rename your files if they have other extensions File Import Unscrambler ASCII Imports files from The Unscrambler version 1 0 2 3e The file format is described in the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler A ppendices Use Flat F
203. e then click anywhere inside the dialog box to validate the change You may also remove a level by selecting it from the list and clicking Remove or change the order of the levels with the Move Up and Move Down buttons Note The order of the levels in the list is important it determines in which order the levels will be stored in The Unscrambler which color will be associated to each level if you use Sample Grouping on a plot etc When you are satisfied click Finish to exit the Category Variable Wizard and go back to the Editor You may then select the proper level of your new category variable for each sample 124 e The Unscrambler Menu Options The Unscrambler Program Operation Category Variable Wizard Select Sets You enter this dialog if you want to base the levels of the category variable on a collection of sample sets Simply select the sets you want in the field Available Sets The selected sets must not be overlapping The category variable will have as many levels as selected sets The levels will take their names fom the selected sets and each sample will automatically get the correct level Samples which are not part of any selected sample set will have a missing value for the category variable When you are satisfied click Finish to exit the Category Variable Wizard and go back to the Editor Edit Insert Mixture Variables This function adds mixture variables to a non designed data table It is not available to
204. e 03 lt Back Next gt Cancel Help Try the two radio buttons and compare the properties of the two designs in order to make up your mind the information displayed gives you the required number of experiments as well as the condition number for the selected design This information is computed from all your former settings in particular it takes into account the Interaction and Square terms that you included in the Define Model dialog Refer to Generate Design D optimal Non mixture p 114 for complementary information on the generation of a D Optimal design in this dialog Generate Design Next Dialog Once you have selected the suitable options and if you chose a D optimal solution only after you have pressed Generate and are satisfied with the result you may press Next The next dialog depends on the specifications of your design e If some of the variables are constrained Last Checks The Unscrambler Program Operation The Design Wizard e 115 e Otherwise Randomization Details General Design Wizard Design Type This applies to Fractional Factorial designs only When you have specified all non design variables you proceed to the dialog Design Type where you specify either the Resolution you wish to achieve or the Number of Experiments you can run If you choose to specify the resolution The Unscrambler will automatically create a design with as few experimental runs as possible If you specify
205. e 11 Camo Software AS The Unscrambler User Manual InstallShield Wizard InstallShield Wizard Complete Setup has finished installing The Unscrambler License Server on your computer Cancel Requesting an Unscrambler License Server Activation Key 10 Start The Unscrambler License Server from menu Start Programs The Unscrambler License Server E PMT G oluc 9 2 e oLuP 9 2 ei Product Optimizer H The Unscrambler 9 5 fg The Unscrambler License Server ROS RR P JB The Unscrambler License Server 9 5 BB pspbrwse jbf 11 The Unscrambler License Server Activation dialog opens Unscrambler base z EE 12344E74 Obtain Unscrambler IT e activation key Activate Currently The Unscrambler v9 6 is not activated Click on the Obtain button to request your Unscrambler activation Cancel e key on lin IF you have no internet connection contact your Camo office and provide us with your Unscrambler machine code 12 e Installation and Configuration The Unscrambler Program Operation If you have Internet access Click the Obtain button to request your Unscrambler License Server activation key online You will be brought to the activation key request form on Camo s website When you submit your request your Unscrambler activation key will be automatically sent to your e mail ID Note Your e mail ID is the means of identifying you as a registered licensee Make su
206. e 48 Options 134 Outlier List 172 PCA 258 PCA Overview 177 Predicted vs Measured 194 Prediction 209 282 283 Prediction Three Way PLS 284 Principal Component Analysis 258 Print 82 Print Lab Report 82 Print Topics 300 Progress 254 Properties 84 244 Randomization Details 117 Reduce Average 233 Regression 268 269 Regression Three Way PLS 273 Regression and Prediction 199 Regression Coefficients 196 Regression Overview 191 192 Regression Overview three way PLS 192 Residuals 186 Response Surface 200 277 Response Surface Overview 202 Results 289 Sample Outliers 181 Sample Properties 244 Sample Statistics 153 Savitzky Golay Derivatives 230 231 Savitzky Golay Smoothing 221 Scaling 155 Scores 183 Scores and Loading Weights 182 Scores and Loadings 181 Select Design to Extend 100 Select Design to Modify 100 Select Design Type 98 Select Existing Design 100 Select Method to Use 97 98 select sample sets to sort 237 Select Sample Sets to Sort 237 Select Samples 132 252 Select Sets 125 select transformations 90 Select Transformations 90 select variable sets to sort 237 Select Variable Sets to Sort 237 Select Variables 133 252 Set Editor 246 Set Weights 262 Shift Variables 238 Sort Samples 235 236 Specify Levels 124 Spectroscopic Transformation 224 SQL Query 58 Standard Normal Variate 232 startup 23 26 Startup 26 Statistics 253 System Setup 90 System Setup supervisor 23 Test Set Validati
207. e Broadcasting Server IP Address SC Iw Create System Log file IV Delete System Log file C Never C Every time Server Shutdown e When size of the file reaches 1 MB Downloading The Unscrambler Client 15 Download The Unscrambler Client installer from www camo com and save it in a shared location on your network 16 Extract the files A folder The Unscrambler Client is created 17 From each client machine double click on file setup exe to start the installation The Unscrambler Program Operation Network Installation Multiple User Licenses e 15 The Unscrambler User Manual Installing The Unscrambler Client If you are upgrading to a new version it is recommended to uninstall your previous Unscrambler installation prior to installing the new one Un installing your former Unscrambler version will remove the program files only all your Unscrambler data files and model files will be preserved 18 If you have downloaded The Unscrambler from Camo s website double click on file setup exe file found in folder The Unscrambler Client to start the installation If you have received The Unscrambler installer on CD ROM place the CD in your CD player 19 AnInstallShield Wizard dialog opens InstallShield Wizard The Unscrambler Setup is preparing the InstallShield Wizard which en will guide you through the program setup process Please wait i Checking Operating System Version 20 The InstallShiel
208. e been extracted from the data or Leverage which shows the leverage of each sample If you choose Sample Residuals you must also choose which sample to look at in the Sample field and in both cases you must choose how many components to use in the prediction You should use the optimal number of components you found for the model based on validation Plot Line General View This command selected from the General Viewer launches a dialog where you can specify a line plot of any vector of your choice from any Unscrambler file Select the file from which you want to plot in the Source field If you do not remember the file name click the Browse button to launch the Results dialog described in Results All There you can search for the right file Press Select to insert the correct file name in the dialog Source File Is A Data Table Once you have selected a source you must specify whether to plot samples or variables in the bottom right field Then select which sample s or variable s to plot either by entering it directly or by pressing Select to enter an Editor where you can mark your selection Ranges are indicated by a hyphen ranges or separate items are separated by a comma Source File Is A Result File In this case you must specify from which matrix stored in the result file you wish to plot Then select which dimension of the matrix you wish to pick elements from the dimension which you have not selected is th
209. e changes into immediate effect The action cannot be canceled 90 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS dp Dialog System Setup System Setup E x General Editor Viewer Missing Data m Miscellaneous IV Confirm deletions IV Kill subwindows on close Most recently used file list Size fa Empty Analyses V Copy data notes to result notes Cancel Help The dialog contains four sheets General Editor Viewer and Missing Data Note If you have logged in as Supervisor the System Setup dialog is different from the current description lookup Chapter System Configuration in Practice for more information System Setup Dialog General Sheet The options in this sheet rule the general use of The Unscrambler in connection to your user account In the Miscellaneous field you can tick one or more of the three available tick boxes e If you tick the box Context Saving The Unscrambler saves the entire workspace when you exit All Editor and Viewer documents will be opened again the next time you open The Unscrambler e The box Confirm deletions ensures that The Unscrambler asks you for confirmation before deleting any samples variables etc e An Editor or Viewer window may have other windows derived from themselves for example a Sample Statistics Editor opened as a sub window Select the
210. e ee eege ESCH 126 Fiditi Pin E OT 127 Edit Correct Mixture Components 2 0 0 eeeeceseeesceseeeeeecseeeceeeeeeeceeesaeeceees caaseseeseeaeeaees 129 Edit Convert to Category Variable 0 cece ceeeseseeeeeecseesceeeceeeceessaeeeseees caaseseesseeeeeaees 129 Edit Convert to Mixture Variable cece cece cesses ceeeseeneeeesseeeesesseeseeeesecseeseeeesaeeaeeaees 131 Edit Split Category Variable oe cceeseseeeeseeseceeeesecaeeaesecesceees cuaeeeseessaeeeseesaeeeseees 131 Edit GoTo ET 131 Edit Select Samples eege dee EEN deeg EE 132 Edit Select Variables ss cvs cssscssvcccsvesenecacsevass covtonsestecevtokaestenenqssvensedtessepsdenssssbapunaneseerenteseees 133 Edit Select All eene sas cnscaeeseededsovbectessnnstena decode sess dace cosebsosesedeiseueavevsnsbeetesaieys 133 Edit Adjust el EE 133 Fidit Add PlOts a cssssiesesssessdsccsses sctupssetcesetenseesuedeseteauedaseb EENEG 134 Bdit OPtlONs 5 csssrescraccasenssepebdasasaecs anena aE EEEE TRAE AEn EEEIEE EER EENT secs 134 Sample Subgrouping Dialog 137 Options Dialog General Sheet for 2D 3D Scatter And Normal Probability Plots 137 Options Dialog General Sheet for Histograms esesssssseresesserere cneeseeeeeseceeeesesaeeeeeaeees 139 Options Dialog General Sheet for Line and Matrix Piot 139 Options Dialog General Sheet for Predicted with Deviations and Percentiles 141 Options Dialog General Sheet for Mean and SDev oo eee ee ceeeee
211. e mat file where filename represents what you have entered Reading the File in Matlab To load the converted file type load filename in the Matlab command window If the data are exported without sample and variable names the filename mat file contains one variable called Matrix that contains the Unscrambler worksheet data If the data are exported with sample and variable names the filename mat file contains 3 variables Matrix ObjLabels and VarLabels where ObjLabels and VarLabels are character arrays that contain the row and column labels respectively Note The variable names Matrix VarLabels and ObjLabels are the same in each exported Unscrambler file Thus if you want to load several converted Unscrambler files into Matlab rename the variables in Matlab after each load command or they will be overwritten by subsequent import operations Missing Value Conversion Missing values in an Unscrambler worksheet are converted to the number 9 9730e 023 Category Variable Conversion Category Variables are converted into integers The Unscrambler Program Operation The File Menu e 73 Export MVACDF Files In The Unscrambler 7 6 a new export format was implemented to answer the specific needs of one of our major Unscrambler users This format is called MVACDF Upon choosing Type MVACDF Files in the Export dialog you enter the Export MVACDF dialog shown below Export MVACDF dialog Export MYACDF x Sets Sampl
212. e one that will vary in the plot The Unscrambler Program Operation The Plot Menu e 211 For instance in the figure below the source file is a PCA result file called Blue The selected matrix is the scores matrix Tai and PCs number 1 3 are selected which means that three lines will be plotted The plotted elements for the first line are the scores along PC1 for the 11 samples for the second line along PC2 and for the third line along PC3 Dialog Line Plot in General View Line Plot Source Bue Dose EI Le Cancel Matrix Tai 5PCs 11 Samples Scores si PCs Samples Note You may select more than one vector at a time each vector will appear as a separate line on the plot In the bottom right corner of the dialog a preview screen shows how the plot will look in the General View Note The preview plot is a general example and does not show what your actual data look like Once the plot is displayed in the General Viewer you may add more lines with the command Edit Add Plot This opens a new dialog where you can select more vectors to be plotted together with the ones already displayed Plot 2D Scatter General View This option launches the 2D Scatter Plot dialog First you must specify the file which you want to plot as abscissa horizontal axis in the Abscissa field and the one you want to plot as ordinate vertical axis in the Ordinate field If you do not remember the file nam
213. e s click the Browse button in the field s in question to launch the Results dialog described in Results All There you can search for the right file Press Select to insert the correct file name in the 2D Scatter Plot dialog Dialog 2D Scatter Plot in General View 2D Scatter Plot Abscissa Calibdata Browse B C Samples e Variables Sam Set Training 28 1 28 E Ordinate Calibdata E C Samples Variables Sam Set Training 28 1 28 212 e The Unscrambler Menu Options The Unscrambler Program Operation nscrambiler User Manual Camo osortware Source File Is A Data Table Once you have selected a source you must specify whether to plot a selected sample or variable along the abscissa and the ordinate Then select which variable set or sample set to take into account For instance in the dialog example shown in the above figure the same data file Calibdata has been chosen as source for Abscissa and Ordinate The vectors to be plotted against each other are variables number 5 and 7 The samples which will appear as points on the plot are taken from sample set Training Source File Is A Result File In this case you must specify from which matrix stored in the result file you wish to plot your Abscissa vector and from which to plot your Ordinate vector Then select which dimension of the matrix you wish to pick elements from the dimension which you have not selecte
214. e sample names are imported Sample names after import with without sample number Original sample Sample name name ith Include Sample number BWX98 23 BWX98 24 BWY01 25 BWY01 26 BWY02 27 BWY02 28 Note Sample names will only be imported if they are present in the source JCAMP DX file 60 e The Unscrambler Menu Options The Unscrambler Program Operation The field JCAMP File Information at the bottom displays summary information about the last file selected in the file list File Import Tracker Through this option you can import Tracker files which are written by Multispec instruments from Foss Electric Several predefined Sets e g Sets for calibration samples are created automatically during import of the file File Import NSAS You enter the Import dialog where you select the NSAS file you want to import from the list of files in the folder NSAS Data Import will read information in the NSAS data file which has no natural place in The Unscrambler s file format into the Instrument Info block File Properties under specific keywords Similarly NSAS Vision Model Export will look for a relevant subset of these keywords and if found it will place the values in the corresponding places in the NSAS Vision Model file File Import APC This option is used to import files generated by APC a package used with some Infralyzer instruments The data are stored in an observation fil
215. e set D Samples 20 Di Define Cancel Variable set Help Sensory 12 Di Define Global Attributes Data set origin Oso Equipment ID ban Equipment manufacturer if Equipment type oo Operator name oO Experiment date time fo Select the sets that make up the data matrix you want to write to the MVACDF file in the Sets field Define new sets in the Set Editor if the already existing sets do not contain the correct data Then in the field Global Attributes you enter all other relevant details Note If you also collect your data in a particular format and would like The Unscrambler to be compatible to this format please take contact with us File Export Model The Unscrambler models can be exported to several formats See Chapter File Export for more information about how to export data How to Export Models You need to specify the models which are to be exported in the respective dialogs specific to each export format Supported Formats for Export Model The Unscrambler lets you export files as the following formats 74 e The Unscrambler Menu Options The Unscrambler Program Operation ASCII MOD see File Export Model ASCII MOD U5 Model see File Export Model U5 Model NSAS see File Export Model NSAS Vision see File Export Model Vision Tracker Model see File Export Model Tracker File Export Model ASCII MOD
216. e transformations used to correct the baseline of samples and are set in the dialog Transform Baseline They are mostly used for spectroscopic purposes The two transformations can be executed separately or together in the combined case the Linear baseline correction will be run first then the Baseline offset Dialog Transform Baseline Transform Baseline Lx Sope Samples Cancel e Define m Variables he e Define How do you want to transform the baseline Baseline offset I Linear baseline correction Baseline correction variables Ast Var bh 1100 nm 2nd Var 130 1300 nm First select the Sample and V ariable sets that define the matrix you want to correct in the Scope field Then select the wanted transformation If you select Linear Baseline Correction you have to select the two variables which define the new baseline The first and last variable are selected by default The Define buttons in the Scope field take you to the Set Editor where you can define new sample and or variable sets 230 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Baseline Offset The formula for the baseline offset correction can be written as follows f x x min X where x is a variable and X denotes all selected variables for this sample For each sample the value of the lowest point in the spectrum is subtracted from all the variables The result of th
217. e values of the levels in the same way as for continuous variables In addition you may Add or Delete levels Delete will remove the selected level and pressing Add will add a new level allowing you to enter the value in the field below the list Click once in the list to update it If you have selected more than one variable in the Editor the properties of the first variable are shown In such a multi selection you can only change the variable name not the levels Modify Layout or Ctrl L Through this option you can change the layout of the data values of selected variables in the Layout dialog The layout settings only affect the visual representation of the data Select the default format of the data values in the Numeric format field The three possibilities are displayed below Use the spin button to select the number of decimal places to display in the data table Values stored with more decimal places are rounded to the number you select The Unscrambler Program Operation The Modify Menu e 243 Numeric format options ption Description The Unscrambler decides the format 12345 6789 Normal layout 1 2345e 04 Scientific notation Note This operation applies to the selected variables only If you wish to change the default layout of all variables in all tables you can use the System Setup dialog Modify Edit Set In the Set Editor dialog you can define sets to be used for data selection
218. e values within a segment of data points 2 Savitzky Golay finds a data value by making a polynomial to fit the data points using a number of data points on each side 3 Median Filter finds a data value by taking the median within a segment of data points 4 Gaussian Filter finds a data value by computing a weighted moving average within a segment of data points Modify Transform Smoothing Moving Average In the Moving Average Smoothing dialog you can replace each value in a row by the average of its nearest neighbors 220 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS First select the sample and variable sets that define the matrix you want to smooth in the Scope field The Define buttons in the Scope field launches the Set Editor where you can define new sample and or variable sets Then you enter the size of the segment to be averaged i e how many adjacent columns should be used to compute the average value in the Parameters field Modify Transform Smoothing Savitzky Golay Savitzky Golay is an averaging algorithm that fits a polynomial to the data points The value to be averaged is then predicted from this polynomial equation Dialog Savitzky Golay Smoothing x Samples Cancel Spectra 100 Define emca H Variables Sa Spectra 100 Define Parameters Smoothing points Number of smoothing points
219. e variables with the mouse Note The sample grouping information is more straightforward to read if you select the same variable in the Group By field as in the Markers Layout field However grouping by one variable and displaying the values of another variable as markers may be useful to include more information into a single plot Sample Subgrouping Dialog In the Sample Subgrouping dialog select which design variable s must be used to generate sample groups for the Sample Grouping display Dialog Sample Subgrouping Sample Subgrouping EN Group on Variables Available Variables i Morpholine a Cancel lt lt lt Help gt gt gt N Number of Sub groups 5 This dialog consists of two fields Group on Variables and Available Variables Select the variables you wish to use from the latter field and move them to the former using the appropriate arrow button The number of sub groups is displayed at the bottom of the dialog the number depends on the variables numbers of levels Choosing one design variable with three levels e g will give three sub groups while two such variables will give 3 x 3 9 sub groups Options Dialog General Sheet for 2D 3D Scatter And Normal Probability Plots Accessed from Edit Options for a 2D scatter plot 3D scatter plot or Normal Probability plot this sheet lets you define the plot layout The Unscrambler Program Operation The Edit Menu e 137 Dialog Options General
220. e way PLS results this dialog allows you to produce a quadruple plot of the main results namely scores X and Y loading weights residual Y variance validation and predicted versus measured Y value as shown in the Plot field in the Regression Overview dialog 192 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Regression Overview for Three Way PLS results Regression Overview OK Eee Cancel Plat Scores Loading Weights Residual Y Validation Variance Predicted vs Measured variables Primary X C Secondary x Y variable 1 Severity si Components Suggested no is 4 Scores and Loading Weights Vector 1 f1 H Vector 2 2 gt Predicted vs Measured 4 e Inthe X variables field choose whether to display Primary or Secondary X variables e In the Y variable field choose which Y variable to study e Inthe Components field choose the PCs to be used as axes for Scores and Loading Weights and the number of components for Predicted versus Measured It is recommended to use the optimal number of PCs for plotting Predicted vs Measured Plot X Y Relation Outliers This is a plot of t scores vs u scores which can be used to detect outliers in PLS The samples should lie as close to each other as possible along a straight line through most of the samples Samples that stick out from this line are possible outliers The plot is specified in the dialog X Y Relation Outliers where
221. ect whether to see Primary or Secondary X variables Plot Scores The score plot shows the similarities and differences among the samples allowing you to investigate patterns and look for outliers The Scores dialog offers three different layouts for plotting each available from its own sheet 1 General one plot window 2 2x2D Scatter two windows 3 4x2D Scatter four windows The Unscrambler Program Operation The Plot Menu e 183 Dialog Scores Scores General 2 2D Scatter 4 2D Scatter OK Cancel rl Help Plot Type C Line 2D Scatter C 3D Scatter m Components Suggested no of components is 1 Vector 1 fl 2 r Samples Vector 2 2 IV Calibration Vector 3 E a M Validation Scores Dialog General Sheet On the General sheet in the Scores dialog you specify one plot You must decide whether you want a Line 2D Scatter or 3D Scatter plot of the scores in the field Plot Type and specify which components you want to see in the Components field Use the default PCs 1 and 2 unless you have specific reasons for doing otherwise Finally make your choice in the Samples field Calibration means that the scores from the calibration samples i e the samples used to construct the model will be plotted If you have used a test set validation in the computations you can also choose Validation which will give you the scores for the validation samples The pr
222. ed to samples Vars he variable number for all positive tests that are related variables is given in this column OW ach test has its own warning number so that you can see which warnings are generated from the same test Description his column describes the test that was positive imit his is the limit that the test value is compared to The limit can be changed in the Warning Limits dialog You can access this dialog from each dialog where you make a model EN 172 e The Unscrambler Menu Options The Unscrambler Program Operation View Toolbars This option lets you toggle the toolbars on and off and is by default always enabled The toolbars contain buttons giving you quick access to the most frequently used commands You select which toolbars to display by selecting the appropriate check boxes in the Viewer Toolbars dialog The toolbars can be moved to different locations Click on the toolbar background with the left mouse button and drag it away The toolbar is then floating and you can place it wherever you want on your desktop also outside The Unscrambler View Status Bar This option is enabled by default and toggles the status bar which gives a lot of information on and off A check mark appears to the left of this command when the status bar is turned on The left area of the status bar describes actions of menu items as you use the arrow keys to navigate through menus and shows messages
223. ee ceeeeeeeneceeeeeeeenes 142 Edit Insert Draw M i i ea E ssensseesosensser E sees 143 Edit Insert Draw Item Line srice aa EOV 144 Bdit lt Inse rt Draw Item VOX a tessecssssssisesceesstsuadsasencetecespenese spores EPAPER EESE R SE EEF EErEE 144 Edit Delete E 144 elleng PrOpe ENE PEENTE dd eg Ebene ebe 144 Edit Math wsscicscsscesucossessstseceuseccncoseustenceseedecnsosevsneneustecesssecevneshcecnsseatsscesbostuessavansesveessaanseasees 146 Edits Mark Outliers Only sic 3cs s5cctaschsstshasnssscdssesvevasesnagedesdfaschacsctaseeshisesehsivacabeposnapapepaesaess 146 Edit Mark Test Samples Only AEN EEN coaeceeseeseeseaeesseeags 146 Edit Mark Evenly Distributed Samples Only 147 Edit Mark Significant X Variables Only 147 Edit Mark With Rectangle AA 148 Edit Mark One By One oo eee ccessee ceeeeeeaceeseecsaeeeseeesaesevaesecseseesaeeaeeeesaceaseaeeaseanes 148 Edit Mark Unmark AU ccs ssstvsccsivees cases satevesensnsetcte n A ted eacnstvecusenetseeswsnaeveiviess 148 Edit Mark Reverse Making 148 tt COpy As eege ege 149 Bt Select Rate ss esses sseves cazstesrsss tages ss eege deet e are e Ee E A a EE 149 Choose Significance Level cic sgeeeiueg geed Cegegeeen th E E E EE o 149 The View MCN 2 tetdete eege ege ER deteeces e RTAS RIE AES ETETE 151 vi e Contents The Unscrambler Program Operation Kal 151 View Variable Stati Sts a 26c0csaseaci hasta cactaches east EErEE OES AE E AE OR AE E ENE
224. ee way Data Add Start Noise This option is disabled for Three way PLS models Issue Warnings Finally The Unscrambler can give you warnings for several situations which may indicate poor validity of the results For example the program can detect high leverages outliers in sample and variable space unusually low total explained variance etc For Three way PLS models the Issue Warnings box is ticked by default and this option cannot be changed 276 e The Unscrambler Menu Options The Unscrambler Program Operation Task Response Surface The Response Surface dialog is a dialog for polynomial regression analysis On the three sheets in this dialog you set up which data the model will be based on and how complex it will be The response surface gives a graphical picture of how the response Y depends on the predictor variables X their interactions and squares It can be used for optimization e g to find the settings of the predictor variables that give the optimal response For non designed data you can select subsets of samples to omit and decide which X variables and Y variables to use For designed data you can only choose which Y variables to incorporate The model can have any degree of complexity from purely linear in the X variables to full quadratic includes all interaction and square effects The dialog contains three sheets Samples X variables Y variables as well as a Warnings field at the bottom Dial
225. el in the Model field Discrimination 208 e The Unscrambler Menu Options The Unscrambler Program Operation Power and Classification plots involve two models select the second model from the drop down menu at the bottom Dialog Classification Classification Plot Coomans Sivs Hi C Si SO vs Hi Classification Table C Modeling Power Model Data From Model Setosa e Projected Onto Model Versicolor S Plot Prediction The Prediction dialog lets you plot the predicted values from a prediction or the residuals and leverage of the predicted samples from two different sheets Predicted and Residuals Leverage Prediction Dialog Predicted Sheet On this sheet you can define a plot of the predicted Y values from a regression analysis The Unscrambler Program Operation The Plot Menu e 209 The Unsc Prediction Dialog Predicted Sheet x Predicted Residuals Leverage m Plot Ce Predicted with Deviation Predicted vs Reference IV Include Table Y variable fi Print through r Components Suggested no of components is 2 p z You can choose between two different plots in the Plot field 1 Predicted with Deviation which shows the predicted values with an estimate of the uncertainty deviation connected with the prediction of each sample 2 Predicted vs Reference which plots the predicted values versus the reference values The
226. elow Here are three basic rules regarding Mixture Designs 1 In a Mixture Design with additional process variables process variable squares may be included only if your continuous process variables have more than two levels 2 Aninteraction between a process variable and a mixture variable can only be added together with the set of all interactions between the same process variable and every mixture variables 3 Interaction and square effects among mixture variables can only be added as a complete set that is all the I amp S effects together can be added removed from the variable set no individual selection is allowed Interaction and Square Effects Selection In this dialog you can select the interactions and squares that you wish to include in the model without taking a whole predefined set The dialog box contains two lists Selected Effects to the left and Available Effects to the right The latter lists all available effects with their full names Select the ones you wish to include and press the Add button to include them in the leftmost list under short names You can also Add All Add Int i e Interaction Effects and Add Sau i e Square Effects Use the Remove or Remove All buttons to remove effects from the leftmost list If your Variable Set contains more than 64 variables you must push the Select sub set button to define a set with max 64 variables before you can select effects Dialog Interaction and Square Effects
227. ence the estimation of the components This is often used if the variables 1 are measured with different units 2 have different ranges 3 are of different types Sensory data are sometimes standardized if the scales of the different attributes vary greatly Spectra are usually not standardized Note If noisy variables with small standard deviation are standardized their influence is exaggerated which can sometimes make the model less reliable The Unscrambler Program Operation The Task Menu e 263 Using constants other than A 1 0 and B 0 0 gives you the possibility to weight the variables with a combination of 1 SDev and a constant This can be a useful compromise in cases where full standardization might be dangerous 1 0 1 0 represents no weighting at all i e all variables are used as they are Constant Constant sets the weight to a constant number independently of the standard deviation Passify Passify sets the weight to a very small constant number so that the variable loses all influence on the model but still appears in projection plots Once you click Update the variable appears as Passified in the list Cross Validation Setup Dialog If you choose the method Cross Validation in a dialog the Setup button next to it will be enabled Press this and you enter the Cross Validation Setup dialog where you can choose which cross validation method you want to use and the properties of that method Dialog
228. ense agreement is a legal agreement between you the Licensee and CAMO PROCESS AS the Licenser and the company you bought the software from for the CAMO software product The software products covered by this agreement is The Unscrambler R T he Unscrambler A Designer The Unscrambler R MYA and The Unscrambler License Server hereafter referred to as the Program and includes computer software and associated media and printed materials and may include on line or electronic documentation By installing copying or otherwise using the Program you agree to be bounds by the terms in this license agreement If you do not agree to the terms in this license agreement promptly return the unused software to the place from which you Do you accept all the terms of the preceding License Agreement If you choose No the setup will close To install The Unscrambler License Server you must accept this agreement InstallShield lt Back 7 Select a destination folder for The Unscrambler License Server 10 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual Camo Softw InstallShield Wizard Choose Destination Location Select folder where Setup will install files InstallShield 8 The program is being installed InstallShield Wizard Setup Status 9 When the setup is complete click Finish The Unscrambler Program Operation Network Installation Multiple User Licenses
229. er Licenses e 13 UnscrambLicServer E P Serial Key Successfully Added 14 The Users Status Dialog will appear Users Status Dialog Server Administrator Help Total number of user fe i Number of Simultaneous Users f Cancel Number of Users Connected 0 The dialog box indicates the total number of users and the number of simultaneous users authorized by your network license Defining User Names and Passwords Use menu Administrator Users in the Users Status Dialog to access the dialog box for user names and password definition Passwords must be between 6 and 15 characters in length and must be a combination of numbers and letters 14 e Installation and Configuration The Unscrambler Program Operation The Unscrambler License Server Henri de Rodez Linda Hanssen Maria Falcao Delete Martin Kermit Oliver Tross i Trond Amundsen Modify Val rie Lengard Details Close Note In the New User dialog box it is important not to check the Can empty log option if your company is submitted to traceability regulations and 21 CFR Part 11 FDA guidelines for electronic recording This option would allow the user to delete the Unscrambler s log trail of data files changes and model files generation Configuring Settings Use menu Administrator Options in the Users Status Dialog to open the Configuration Settings dialog where you may configure the system log file Configuration Settings 3 Iw Enabl
230. er Manual p 208 209 Click OK to access the next dialog Import 3 D where you will specify your 3 D table layout see The Import 3 D Dialog for a detailed description File Import 3 D Excel The Import Worksheet dialog allows you to specify the parameters of the import from an Excel source file It is described in detail in The Unscrambler User Manual p 210 Click OK to access the next dialog Import 3 D where you will specify your 3 D table layout see The Import 3 D Dialog for a detailed description 68 e The Unscrambler Menu Options The Unscrambler Program Operation File Import 3 D JCAMP DxX For JCAMP Dx files the standard file browser is replaced by the dialog JCAMP DX Import where you may select one or several source files for the import No additional parameters need to be specified since all relevant information is stored with the JCAMP DX format This dialog is described in detail in File Import JCAMP DX Click OK to access the next dialog Import 3 D where you will specify your 3 D table layout see The Import 3 D Dialog for a detailed description File Import 3 D Matlab The Import Matlab dialog allows you to specify the parameters of the import from a Matlab source file version 5 x up to 7 0 It is described in detail in File Import Matlab Note The Matlab import routine has been re written completely to accommodate new features related to sample and variable names or 3 way data s
231. er Menu Options The Unscrambler Program Operation Dialog Statistics Group sheet Statistics General Compressed Group Scatter i Cancel Help m Plot Mean and Standard Deviation Samples Plot no Groups defined by 1 Ja D Number of selected Groups 12 Decide which variable to use for sample grouping in each plot in the Samples field The mean and standard deviation are computed for each level of the variables Statistics Dialog Scatter Sheet The plot you specify on this sheet is only relevant for Spectra It shows the selected samples plotted against the average sample Scatter effects are present if the slopes of the lines are different or if the baseline is changing from sample to sample and are caused by e g different size of particles opaque solutions etc They can be removed with MSCorrection or differentiation You only need to specify which samples are to be plotted this way in the Samples field Statistics Dialog Cross Correlation Sheet This sheet allows you to display the cross correlations among all variables in the analysis either as a matrix plot or a table or both By default both boxes are ticked in the Plot Type field If you wish to display only the matrix plot un tick the Table box and vice versa When displaying only one plot you may double click on the Plot Preview to select a single subview viewer The Unscrambler Program Operation The Plot Me
232. eral View Note The preview plot is a general example and does not show what your actual data look like Plot Matrix General View This command selected from the General Viewer launches a dialog where you can specify a plot of any matrix of your choice from any Unscrambler file Select the file from which you want to plot in the Source field If you do not remember the file name click the Browse button to launch the Results dialog described in Results All There you can search for the right file Press Select to insert the correct file name in the dialog Source File Is A Data Table Once you have selected a source you must specify whether to plot samples or variables in the bottom right field Then select which sample s or variable s to plot either by entering it directly or by pressing Select to enter an Editor where you can mark your selection Ranges are indicated by a hyphen ranges or separate items are separated by a comma Source File Is A Result File In this case you must specify which matrix stored in the result file you wish to plot Note1 Only one matrix can be plotted at a time Note2 Only a whole matrix can be plotted In the bottom right corner of the dialog a preview screen shows how the plot will look in the General View Note The preview plot is a general example and does not show what your actual data look like Plot Normal Probability General View This command selected from the Gen
233. eral Viewer launches a dialog where you can specify a normal probability plot of any vector of your choice from any Unscrambler file Select the file from which you want to plot in the Source field The Unscrambler Program Operation The Plot Menu e 215 If you do not remember the file name click the Browse button to launch the Results dialog described in Results All There you can search for the right file Press Select to insert the correct file name in the dialog Source File Is A Data Table Once you have selected a source you must specify whether to plot samples or variables in the bottom right field Then select which sample s or variable s to plot either by entering it directly or by pressing Select to enter an Editor where you can mark your selection Ranges are indicated by a hyphen ranges or separate items are separated by a comma Source File Is A Result File In this case you must specify from which matrix stored in the result file you wish to plot Then select which dimension of the matrix you wish to pick elements from the dimension which you have not selected is the one that will vary in the plot For instance in the figure below the source file is a PCA result file called Blue The selected matrix is the scores matrix Tai and PC number 1 is selected which means that the plotted elements are the scores along PC1 for the 11 samples Dialog Normal Probability Plot in General View Normal Probabil
234. ert a row or a column of a data table 2 D or 3 D into a two way data table This feature is particularly useful when the selected vector is one of the following e Column of 3 D data table with layout ON e Row of 3 D data table with layout ov e Column of a data table containing unfolded three way data with several blocks below each other e Row of a data table containing unfolded three way data with several blocks side by side Once the vector has been converted into a 2 D data table you may perform operations which were not possible with a 3 D layout Note Only one vector at a time can be converted The figure below shows the Convert Vector to Data Table dialog Dialog Convert Vector to Data Table Convert Vector to Data Table EN Select which vector to convert then how you want this vector mapped into a 2 D data table Cancel r Vector ot a i dili r Size of new data table Variables j5 4 Samples E In the Vector section choose an orientation Row Column and select the desired vector In the Size of new data table section tune the numbers of Variables and Samples Click OK to display the new table File Duplicate You can duplicate designed data tables in The Unscrambler There are 4 options under File Duplicate e Duplicate as Non design e Duplicate as Modified design e Duplicate as 2 D data table e Duplicate as 3 D data table 80 e The Unscrambler Menu Options The Uns
235. es Name qualifier Numeric qualifier Double quote C Single quote None C Double quote C Single quote None Item delimiters T Comma IT Semicolon I Space I Other s l String representation of missing data fim If sample and or variable names occur in the first column or the first row you can also import those by selecting the box in the File format section Text or numbers may occur within single or double quotes in the file you import from In the fields Name qualifier and Numeric qualifier respectively you should specify the quotes that are used to make the import more secure However it is now also possible to import names without any quotes even if those names may contain some spaces This applies both to sample variable names and category variable levels Note If names are not enclosed in quotes in your ASCII file they should not contain any spaces if you select space as item delimiter see Item delimiters below Numbers may be delimited by different characters in different ASCII files Specify which delimiter is used in the file you want to import using the field Item delimiters You may choose Comma Semicolon Space or specify another delimiter e g WV Note Carriage Return Line Feed and Tabulation are not among the available delimiters in the dialog They are default item delimiters and will automatically be recognized as such Do not specify them in the
236. es are used only as a comparison to assess the quality of the predictions not to guide the predictions The Unscrambler Program Operation The Task Menu e 283 Prediction Dialog Pretreat Vars Sheet This sheet is useful if you want to perform automatic pretreatments at prediction Note that this option is only available if your model has been calibrated on data pretreated in the Unscrambler Once you have selected relevant automatic pretreatments using the Pretreat button see Prediction Dialog General Options specify which Variable Set to pretreat either by choosing an existing Variable Set from the drop down list or by defining a new in the Set Editor which you enter by clicking Define Note You are free to select a Variable Set to pretreat which differs from the X Variables set But make sure that it contains the same number of variables as the original set of variables pretreated prior to building the model If not you will get an error message as shown below Dialog Invalid number of Pretreatment Variables The Unscrambler P Number of variables selected in Pretreatvar page should be equal to number of pretreatment variables in Model Prediction Dialog General Options You select the model to use in the prediction by entering the model name in the Model Name field or by pressing the Find button next to it to enter the Get Model dialog The number of X variables in the prediction data set must be the same as the number
237. es contact support camo com or Cancel support camo no Login Dialog This dialog pops up once you have chosen a user name from the Startup dialog if the chosen user has a password and the system Supervisor has activated all passwords 26 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog Login here as Supervisor Use Supervisor Cancel Password Help ai Type in your password and click OK To remove the password you have to log in as supervisor see System Configuration in Practice The Unscrambler Program Operation System Configuration e 27 The Unscrambler User Manual Camo Software AS General User Interface Beginners get your marks in The Unscrambler The purpose of this chapter is to give you a global introduction to the principles at use in The Unscrambler In what follows you will be provided with a short overview of the Unscrambler user interface or workplace Working with Windows The Unscrambler runs under Windows 95 Windows 98 Windows NT 3 51 or later Windows 2000 and Windows XP We will assume that you are already familiar with the operating system you are using If not we recommend that you become fully acquainted with it before starting to work with The Unscrambler This manual does not explain in detail important points such as the Windows graphical interface common use of the mouse etc Refer to you
238. es the relationship between one Y variable and a selection of two X variables as a contour or a landscape plot You see how the response Y varies as a function of the selected X variables all other variables are kept at constant levels which are the mean levels by default Three X variables in the case of mixtures The Contour Plot shows the contours or level curves of the surface The optimum value will be indicated if it is within the experimental region The Landscape Plot shows the surface in a three dimensional plot with the X variables in the horizontal plane and the fitted response in the vertical direction All response surfaces plotted in the Contour plot layout display values for the Y variable along the contour lines These values are expressed in the unit of the Y variable They come in addition to the use of several colors they are of major interest in particular if you wish to print out a contour plot on a black and white printer Contour Plot of a D Optimal Design Values are displayed along the contour lines 21 783 29 706 37 629 45 552 53 474 61 397 S x Ge Milk Fat D Response Surface e i i i _Lemnerature A 80 85 90 95 100 105 Pancakes Process PC 3 Y var DLC X var value pH B 6 2265 L C 1 4583 Note All contour plots display values along the contour lines regardless of the type of data plotted The Unscrambler Program Operation The Plot Menu e 199 Pl
239. escription Grid Displays a grid allowing you to see the data points coordinates more clearly Plot Header Draws the title of the plot and axis names In the Labels Layout field you choose how markers and labels will appear in the plot Available options are given below Labels layout options options Description Name elect which part of the name to use in the plot Number rhe basis for the numbering is the current Set ariable Level he level of the sample is used to generate the name for example This Indication Name equires a designed data table Position on file he basis for the numbering is the raw data Orientation field Horizontal or Vertical To the right of the radio button Name you choose how you want to build up the name character string The first mouse click on a character in the name string indicates the start of a character sequence The next click indicates the end of the name string All characters in between are selected If you want to select characters that are not in sequence hold down the lt Ctrl gt key when you select the characters you want in the label with the mouse Deselect one by using lt Ctrl gt and clicking again Options Dialog General Sheet for Mean and SDev Accessed from Edit Options for a special plot Mean and SDev this sheet lets you define the plot layout Dialog Options General sheet for Mean and SDev Options General OK L e Cancel
240. eseceessesaeeeeeae seneeeesaesneseasaeents 113 Design Wizard Design Type 00 0 ccececeeeeeseeseeesseeseceseesecaessesaeeeceees KENEEN EEE Ser EEIE 116 Design Wizard Design Details 116 Design Wizard Randomization Details General 117 Design Wizard Last E 118 The Unscrambler Program Operation Contents e v The Edit Meussen A TEE EEE EES 119 Edits LEE 119 Edit Com cscs cscevsssncscedstencoceussedhcpsesesencesesobsuposerssoncesestsptosedeoatevueshoas TENTEN ETETE RONE ES 119 Feit e 119 Edit Pastene deer e EE 119 Edit UE 120 Blehen 120 Edit Insert E 122 et e Eu EE 123 Bt dng Variable iscsi e nesvigsavuctusscis eeneg 123 Edit Insert Sample or Variable EEN ee ceeceeeeseessesescsesesessesssceeseeseeesseesseseaes 123 Edit Insert Category Variable 000 ice ececeeeeesseeceesseeeseeesaeeeseessaeeeseees caaeeeseesaeeaeeaeeaseaes 124 Edit Insert Mixture Variables A 125 Bit Append E 125 tt Append Samplesiiscs cicscesisecesccscsansseossstareeoanensFaccessvapspvagancsseestcecensotnscsnscnvsessenecsesseesen 125 Bit Append Variables ecsccsiisss covctisessestbasssdavsrcesencscssssessvecsescadecesseetteenesdteedevtessessbevolesectians 125 Edit Append Samples or Variables AA 125 Edit Append Category Variable wo cece cece cseesseceeseeneesesseeseeaessesaeeae secseeaeeeesaseaeeaees 126 Edit Append Mixture Variables cece cece eseesseceeeeeneesesseeeesesaeeaeeae sesseeaeeeesaseaeeaees 126 Blees E
241. essing the Select button next to any of the two last fields will launch the Select 226 e The Unscrambler Menu Options The Unscrambler Program Operation Samples Select Variables dialog respectively which displays an Editor where you can make your selection directly MSC EMSC Background information setup Dialog This dialog is accessed by clicking Setup after you chose Spectral background information in the Multiplicative Scatter Correction dialog It allows you to provide the EMSC model with prior knowledge about your data Dialog Background information setup Background Information setup Optional EMSC inputs Reference Spectrum Select Good Spectra Subtraction weight Select ser Bad Spectra ti i iSCS Channel weights OK All1 0 Weights Oox Cancel Number of reweightings 0 ia elp The Reference Spectrum field gives you the opportunity to select a single spectrum from your data acting as a typical spectrum without any additional effects If not selected a reference will be calculated using the mean of all spectra In the Good Spectra and the Bad Spectra fields you can specify several spectra from your data table that you believe are good and bad representatives of your spectral data respectively If you have selected data in the Good Spectra and the Bad Spectra fields you are also allowed to enter a subtraction weight for the respective spectra These subtraction weights are multiplied to the
242. essos ngessiadnsotaadecs ue 76 File Export Mod l Vision 5 sccssihverises esterase a ER es 79 File Export Model Tracki eege cescacecessteauresctsenesdsvayuveusiesevyvepsecesensevs NEENA EE 79 File Convert Vector to Data Table oo ccc cccscecseeeseeeeeeeeecneecseesaecaeseeesees seseseeeesessaeeaeaegs 80 Pile Duplicate a scvsceschestessisiencsccvsbiavens E rE EE NE E E EE AREN EEN 80 File Duplicate As Non design ss iisccescescsesssececsusessessacenebesyasshascsbasenestsasdisuasetaseaessafidesdoesrese 81 File Duplicate As Modified Design 00 0 ccc cseeseeecseeeseeeeeseeeeeeesaeeeseeesaeeeseeseeseeeesensaeenes 81 File Duplicate As 2 D Data Table cece ENEE 81 File Duplicate As 3 D Data Table 81 File Print Lab Report enges ees tee ea a Eed 82 SU TEE 82 SUE 83 SUE D 83 SUE EE 83 Fle MOCK EE 84 File Properties eege EE 84 Select Transformations Dialog ss scciessciscescdessasteeschiase chetscsusesehsesasssnssssudescsesabasdsedseneynseesabesastsess 90 Pile Syste Setup gege desedtdescusadevecoscossssbussadevecesevedsecadseusacuscusdsobtestevssenesseeetes 90 File Recent Piles Lasts ecn ora roerien n A R EE actos 95 Pile E LEE 96 The Design WAZA EE 97 File New Desen EE NEEN SEA SEEENER ENEE 97 Design Wizard Select Method To Use oo eeececeseeeseesesseeesseesceeecaeeceeesaeseseeeerseeeeaeeaseas 98 Design Wizard Select Design Type ceeecsceeecseeeeseesesseesceeesaeeeceecsaeesseeeeeseeeeeaeseeeeeaees 98 Design Wizar
243. esults This dialog sheet enables you to select either Primary X X1 variances Secondary X X2 variances or Y variances and display them as a single plot Variance is plotted against component number You can choose between calibration and validation samples or both Explained variance is selected by default you can later change that by using View Source e In the field Variables select which variables you want to display in the plot You can choose between variance for the Primary X variables the Secondary X variables or the Y variables To select All variables click this button Note that the Select button is disabled for three way data Tick the Total box to display the total variance at the same time e You can decide whether you want the variance to be plotted for calibration or validation samples respectively or both by ticking the appropriate box es in the Samples field Calibration variance measures the fit of the model to the calibration data The validation variance is an estimate of the fit of the model to new data not present in the calibration and is based on the validation Plot both measures to obtain full information about model fit The dialog displays a preview screen Double click on this to change the size of the plot Variances and RMSEP Dialog X and Y variance Sheet This dialog sheet enables you to plot X and Y variances simultaneously as a double plot In the field Variables you select which variables
244. eview screen on the sheet lets you switch between displaying the plot in the upper left corner of the screen and on the entire screen Double click in the preview screen to switch between the options Scores Dialog 2 x 2D Scatter Sheet This sheet lets you specify two 2D Scatter plots which are displayed horizontally The number of components must be selected for each sub window Specify which components you want to see in the Components field Then make your choice in the Samples field Calibration means that the scores from the calibration samples i e the samples used to construct the model will be plotted If you have used a test set validation in the computations you can also choose Validation which will give you the scores for the validation samples Scores Dialog 4 x 2D Scatter Sheet This sheet lets you specify four 2D Scatter plots The number of components must be selected for each sub window Specify which components you want to see in the Components field Then make your choice in the Samples field Calibration means that the scores from the calibration samples i e the samples used to 184 e The Unscrambler Menu Options The Unscrambler Program Operation construct the model will be plotted If you have used a test set validation in the computations you can also choose Validation which will give you the scores for the validation samples Plot Loadings The loading plot shows which variables are important and which va
245. ew table of type 2 D data Otherwise you may drag and drop selections between 2 D and 3 D tables provided that the dimensions are consistent For instance dragging and dropping rows into an OV table is OK but dropping columns into an OV table is not allowed because they would not fit into the Primary Secondary variable structure File Import 3 D Unscrambler Results The Import from Result dialog allows you to choose one or several result matrices to be stored as three way data The figure below shows an example where the Eix X residuals matrix has been selected from a Regression model with 103 samples Rows 15 variables Cols and 6 PCs Plns standing for Planes Dialog Import from Result Import from Regression Result x Name Paper 103 Combine Rows C Columns Cancel Matrix name Rows Cols Pins jans BER B BO 1 Bow 1 1 15 Bdiagnosis 5 6 BdiagnosisW 5 Bw ExpxCalS amp ExpxCalT ot Le PCs 0 5 05 IV Insert row and column names Names T Add matrix name to row name J Add matriz name to column name Current selection Number of rows 618 Number of columns 15 Total number of elements 9270 The text on top of the dialog box states the name of the Unscrambler result file you are importing from On the figure above the file name is Paper 103 Below choose whether to Combine Rows or Columns This option plays a double
246. ewer window where the results will be displayed Results MSC EMSC This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type MSC You may take advantage of this function to browse through the existing MSC EMSC model files and check the parameters of a given MSC EMSC model Select the file you wish to view and read the file properties including the parameters of the MSC transformation from the Information field You may press Print to print out the file properties Note MSC EMSC model files cannot be opened in a Viewer so clicking View will result in an error message The Unscrambler Program Operation The Results Menu e 291 Results Data Table This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type All Data You may take advantage of this function to browse through the existing data files and check their properties print them out or print out a Lab Report However data files cannot be opened in a Viewer so clicking View will result in an error message Results General View This command launches an empty Viewer where you can make general plots such as a line plots matrix plots etc This option gives you full flexibility in pl
247. experiments in sequence For example the temperature may be very difficult to change so you want to make all experiments with the same temperature at the same occasion In the dialog Randomization Details which you enter when you click Next in the Design Details dialog you can specify blocks of similar samples to be kept together during randomization Your first option is to Keep reference samples together Alternatively you can select certain variables which are not to be randomized by selecting the appropriate variables in the section Available Variables and copying them to the Do Not Randomize section by clicking lt lt lt or gt gt gt The Unscrambler Program Operation The Design Wizard e 117 Randomization Details Next Dialog Once you have made your selection and clicked Next you access the last dialog of the Design Wizard Last Checks Design Wizard Last Checks This is the last dialog which gives you a summary of your design now that it has been completely specified Click Preview to see the randomized list of experiments If you are not satisfied with the randomization you can use Re randomize If the design is as you want is you may print a lab report by clicking Lab Report Click Finish to end the design specification or Back to review and alter your previous choices Clicking Finish makes you exit the design Wizard and displays the designed data table in the Editor The design can then only be changed by wa
248. f relevant type appear in the list Select file and click Import The Import 3 D Dialog Dialog Import 3 D Import 3 D You have just selected a 2 D import data source with OK 6 rows and 25 columns Now select how to map these data into a 3 D data table ES H m Layout Help O objects x variables x variables C O objects x objects x variables r Size Primary variables 25 Secondary variables fi Samples je Secondary samples fi 4 In the Import 3 D dialog no matter which source you imported from you can specify the 3 D layout of your choice Choose between layouts OV and ON then select the relevant numbers of samples and variables of each type Primary Secondary Supported Formats for Import 3 D The Unscrambler lets you import data from the following formats e Unscrambler Results see File Import 3 D Unscrambler Results e Flat Flat wide ASCII files see File Import 3 D ASCII e Excel files see File Import 3 D Excel e JCAMP DX files see File Import 3 D JCAMP DxX e Matlab files see File Import 3 D Matlab e F3D files see File Import 3 D F3D 66 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Drag n Drop 3 D A selection from a 3 D data table can be dragged in the usual way If you choose to drop it into an empty spot of the workspace this will create a n
249. f the design variables in experimental plans based on two levels of each variable 304 e Glossary of Terms The Unscrambler Program Operation In Box Behnken designs all samples which are a combination of high or low levels of some design variables and center level of others are also referred to as cube samples Design Def Model In The Unscrambler predefined set of variables interactions and squares available for multivariate analyses on Mixture and D optimal data tables This set is defined accordingly to the I amp S terms included in the model when building the design Define Model dialog Design Variable Experimental factor for which the variations are controlled in an experimental design Edge Center Point In D optimal and Mixture designs the edge center points are positioned in the center of the edges of the experimental region End Point In an axial or a simplex centroid design an end point is positioned at the bottom of the axis of one of the mixture variables and is thus positioned on the side opposite to the axial point Gap One of the parameters of the Gap Segment and Norris Gap derivatives the gap is the length of the interval that separates the two segments that are being averaged Look up Segment for more information Histogram A plot showing the observed distribution of data points The data range is divided into a number of bins i e intervals and the number of data points that fall into each bi
250. fine new variable sets click Define to go to the Set Editor If you want to exclude some of the variables from the analysis type in their numbers in the field Keep Out of Calculation or click Select to choose them directly from an Editor Editor the Select Variables dialog Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10 13 27 Progress Dialog This dialog pops up after you have clicked OK in a Task dialog e g PCA etc For time consuming computations you can see the progress of the computations displayed as colored squares in a horizontal bar at the bottom of the dialog 254 e The Unscrambler Menu Options The Unscrambler Program Operation The Unsc nbler User Manual Came Dialog Progress here in the case of PCA PCA Progress Data table Jam design PC Wam XWalidation Variance II The calculation completed successfully While the computations are going on a Cancel button allows you to stop the computations For projection methods PCA PCR PLS the Progress dialog also displays the relevant residual variance after each PC When the bar is full the computations are finished and two new buttons appear at the bottom Select View to display the results in the Viewer Select Close if you do not want to view the results now A warning dialog pops up prompting you for saving the results file Dialog Do you want to save upon clicking Close in Progress
251. fo and object info capabilities Window Cascade This option arranges all open windows so that their title bars are visible To make one the current window simply click its title bar Window Tile Horizontally This option sizes and arranges all open windows above each other so that all are visible All windows have a horizontal layout wide and short Window Tile Vertically This option sizes and arranges all open windows side by side so that all are visible All windows have a vertical layout tall and thin Window Close Current Ctrl F4 Closes the file displayed in the current Editor or Viewer window If you have made any changes since the last time you saved the file you are prompted to save the contents before it is closed 294 e The Unscrambler Menu Options The Unscrambler Program Operation nscrambler User Manual Camo Software AS Window Close All This option closes all open windows returning you to The Unscrambler s default opening screen If you have made any changes in some of the files displayed in those windows you are prompted to save the contents before the corresponding windows are closed Window Identification This is a toggle option which exists for Viewer windows only Click it to activate an extra dockable window at the bottom of the screen which shows object identification of the data in the active plot Each data source is described various curves plotted in a line plot a
252. formation in each dialog and press Next If you make a mistake you can always go back to a previous dialog by pressing Back According to the type of design you have selected the next dialog will be as follows e Full Factorial Fractional Factorial Plackett Burman Central Composite or Box Behnken Define Design Variables e D optimal Non mixture Define Process Variables e Mixture Define Mixture Variables The Unscrambler Program Operation The Design Wizard e 99 Design Wizard Select Existing Design This dialog lets you re use the specifications of an existing design If you remember the name of the file you want to re use you can enter it directly if not you can click Browse to enter one of the two dialogs Select Design to Modify or Select Design to Extend sce below for details where you can look for the appropriate file Select Design to Modify This dialog is accessed by pressing the Browse button in the Select Existing Design dialog and is identical to the Open File dialog Select the data file you wish to modify and press Open Select Design to Extend This dialog is accessed by pressing the Browse button in the Select Existing Design dialog and is identical to the Open File dialog Select the data file you wish to extend and press Open Select Existing Design Modify Next Dialog According to the type of design you have selected the next dialog will be as follows e Full Factorial Fractional Factorial Pla
253. forming you that the model cannot be viewed The limited number of result matrices in a Minimum model also has some effect on the Export Model functionality For Minimum models this option is only available by selecting a model in the Results dialog and clicking the Export button For a Minimum PCA model only the Tracker format is supported Center Data Ticking the Center Data box makes a model where the data have been centered first i e the average of each variable is subtracted from each of the data values in the variable Centering means that the results can be interpreted in terms of deviations from the average this option is highly recommended You should have a good reason for choosing not to center your data before an analysis The Unscrambler Program Operation The Task Menu e 261 The Unscrambler User Manual Add Start Noise You can select Add Start Noise if you want noise to be added to your data before the analysis is performed This option should only be used in special situations where a special data distribution makes it difficult for the algorithms to find a start direction Issue Warnings Finally The Unscrambler can give you warnings for several situations which may indicate poor validity of the results if you tick the Issue Warnings box For example the program can detect high leverages outliers in sample and variable space unusually low total explained variance etc The warnings are issued in accordance t
254. from center in which you may tune the distance between star points and design center Access to this field is allowed once you have defined at least two design variables Define Design Variables Next Dialog Once you have added all necessary design variables by clicking New or changed their definitions by clicking Properties you may proceed by clicking Next You will then access the next dialog Define Non design Variables Add Design Variable Variable Properties Dialog When you select respectively New or Properties in the Define Design Variables the Define Mixture Variables or the Define Process Variables dialog you enter a dialog called Add Design Variable or Variable Properties respectively In this dialog you give a design variable a name and specify its properties The dialog has varying options depending on the type of design you are building and the type of variable you are adding 104 e The Unscrambler Menu Options The Unscrambler Program Operation Add Design Variable Dialog Factorial or Plackett Burman Design Dialog Add Design Variable for a factorial design Add Design Variable E m Data Type Cancel Ze Continuous C Category Help m Levels First you give the design variable a name in the Name field Each variable is given an ID letter A B which you cannot change as well as the name which you define Next you choose whether the variable can vary continuously withi
255. fy existing 98 Plackett Burman 98 99 set up 99 Design Def model 305 Design Details dialog 116 Design Point 100 Design Type Mixture dialog 112 Design Type dialog 116 design variable 305 design variables 31 Design Wizard 97 dialog 2D Scatter Plot in General View 212 3D Scatter Plot in General View 213 Add Design Variable 104 Add Non design Variables 107 Analysis of Effects 278 Analysis of Variance 202 Append Samples 125 Append Variables 125 Background Information Setup 227 Center and Scale 232 233 Change License 301 Change Viewpoint 156 Classification 208 278 Clustering KMeans 255 Compute 217 Configure 201 Cross Validation Setup 264 Data Link Properties 56 Define Design Purpose 112 Define Design Variables 97 104 Define Mixture Variables 107 Define Model 111 Define Non Design Variables 107 Define Process Variables 109 Design Details 116 Design Type 116 Design Type Mixture 112 Effects 203 Enter Multi Linear Constraints 109 Enter Variable Name and Choose Method 124 Estimated Concentrations 191 Estimated Spectra 191 Evenly Distributed Samples 147 Exact Scaling 156 Export 71 Export ASCII 72 Export ASCHU MOD 75 Export Matlab 72 Export Tracker Model 79 Export U5 Model 76 Export Unscrambler 5 71 Export Unscrambler ASCII 71 Extend 100 Extend a Central Composite Design 102 Extend a Centroid Design 103 Extend a Fractional Factorial Design 101 Extend a Full Factorial Design 100 Extend a Lattice Design
256. g File Import Unscrambler Results before you can write them to disk in another file format Importing Data into The Unscrambler Using Drag n Drop The Drag n drop feature is a very fast way to transfer data between Windows applications You can use Drag n drop to import data into The Unscrambler from applications such as Microsoft Excel Note Data tables can only be transferred as text i e it is not possible to link or embed them in other applications i e you cannot use OLE or DDE to transfer them Transferring Data from Spreadsheets into The Unscrambler The easiest way to transfer data from Excel and Lotus spreadsheets is to have both programs open at the same time and use Drag n drop see section Importing Data into The Unscrambler Using Drag n Drop However there are times when you do not want to open an extra program and still want to import the spreadsheet file directly from disk For instance you may not have Excel on your PC but you would still like to import data from an Excel file that you got from a colleague into The Unscrambler In those cases it is smart to have used the Names function of the spreadsheet program and given the different matrices a name You should also do this for sample and variable names If The Unscrambler detects that you have named some areas of the spreadsheet you can quickly select the name from the range list instead of typing the cell area Do this from the Import from Work
257. g ones Likewise if you combine Columns the first selected matrix determines the number of rows in the resulting table and only matrices with the same number of rows can be selected further Note It is not recommended to import several planes from different matrices simultaneously as this function has not been optimized to give a 3 way layout that makes sense Importing several planes from a matrix Some result matrices e g Eix in the figure previous page consist of several rows columns and planes Usually the planes correspond to varying numbers of PCs in the model For models where Martens Uncertainty Test has been computed some matrices have as many planes as cross validation segments If you combine Rows importing several planes from the same matrix will give you a 3 way data table with an O V layout The original Columns of the matrix determine columns of the new table while the original Rows determine Secondary samples and the Planes determine Primary samples If you combine Columns importing several planes from the same matrix will give you a 3 way data table with an OV layout The original Rows of the matrix determine rows of the new table while the original Columns determine Secondary variables and the Planes determine Primary variables File Import 3 D ASCII The Import ASCII dialog allows you to specify the parameters of the import from an ASCII source file It is described in detail in The Unscrambler Us
258. gives a lot of information about the data in the current plot If the Plot ID is turned on a line at the bottom of the plot displays basic information Toggle the Plot ID on and off using View Plot ID The table below shows some typical ways of identifying plots Plot ID syntax Plot type Typical ID line Explanation Score plot Alcohol X expl 70 14 The results file is Alcohol The explained X variance is 70 2D Scatter Y expl 29 28 for PC 1 and 14 for PC 2 The explained Y variance is 29 for PC 1 and 28 for PC 2 Loading plot Octane PC X expl Y expl The results file is Octane The explained X variance is 70 line 1 70 29 and the explained Y variance is 29 for PC 1 Predicted vs Alcohol Y var PC The results file is Alcohol The predicted vs measured results Measured Methanol 3 Methanol 3 for the Y variable Methanol is plotted using 3 PCs Other information about the plotted data such as data source explanation of colors and symbols etc may also be shown in a separate window using Window Identification These windows are dockable views Use View Plot Statistics to display the most relevant statistical measures Information on each object in the plot can be displayed simply by letting the mouse cursor rest on the object in the plot A brief explanation of the data point then appears Click with the left mouse button to display more detailed information about the data object Point Details
259. good and the bad spectra and subtracted from the corrected spectra It should also be noted that the background spectra available for selection in the Select button are those contained within the variable Scope defined in the Multiplicative Scatter Correction dialog This is due to the fact that spectra used as background spectra must have the same number of variables as the spectra to be transformed It is also recommended that you select background spectra to be different samples from the samples in the selected scope of your data table Overlapping reference good and bad spectra is not allowed A warning message will appear if this happens Example Let us say that your data table consists of several spectra measured on different mixtures of two chemical compounds where the amount of each of the two substances is varying The reference spectrum can be a spectrum measured on a mixture where the two compounds are equally represented Good spectra would then be spectra measured on each compound alone The bad spectra could then be selected as spectra believed to contain additional effects not caused by the chemicals The last entry in the Background information setup dialog is the Channel Weights field Here you can choose to select different weighting of your variables It is also possible to iteratively find better weights than the default choice by entering a number in the Number of reweightings field The EMSC will then be run this
260. gs This is a quadruple plot of scores and loadings from a PCA or regression model You will see the scores and loadings for two different sets of PCs This plot is useful for interpreting the relationship between samples and variables in data There are three different ways to plot Scores and Loadings 1 Two plots 2 Four plots 3 Bi plots These are represented by different sheets in the Scores and Loadings dialog Dialog Scores and Loadings Scores and Loadings x Two Plots Four Plots Bi plots EH CAMO rT Plot Scores Loadings Samples Components M Calibration Suggested no of components is 1 IV Validation Vector 1 H Loading Cc Vector 2 IR El Ex c x and Y Scores and Loadings Dialog Two Plots Sheet This sheet lets you specify two plots a Score plot in sub view 2 and a Loadings plot in sub view 3 e Choose the PCs for which you want to look at the plots in the Components field e In the Samples field select whether to plot Calibration and or Validation samples The Unscrambler Program Operation The Plot Menu e 181 e In the Loadings field select whether to see X or Y loadings or both Scores and Loadings Dialog Four Plots Sheet On this sheet you can specify four plots two Score plots plot 1 and plot 2 in sub views 4 and 5 and two Loadings plots plot 3 and plot 4 in sub views 6 and 7 e Specify four PCs in the Components field the two axes common to
261. h transformations were applied to the data before calibration was performed This is important if exported models are to be used for predictions as spectra used as input to the prediction must undergo exactly the same transformations as the calibration spectra NSAS Export dialog Math Sheet NSAS Export E General Instrument Parameters Math 8 OK Ast NSAS transformation Math type Conect for reference changes gt Cancel Help dd m 2nd NSAS transformation Math type Segment size nm fe a Gap size nm 4 1st derivative v NSAS allows for up to two transformations specified in the fields 1st NSAS transformation and 2nd NSAS transformation respectively The first transformation is used for the sole purpose of indicating whether the spectra has been corrected for reference changes within the instrument or not Choose Math type from the drop down list The second transformation may be any NSAS math type including None Select from the drop down list For some of the math types one or more of the parameters Segment size Gap size or Divisor point will be enabled These will be written to the model file exactly as shown Use the spin buttons to specify them Default values for the transformations are taken from the Instrument Information block in The Unscrambler s result file which is again copied from The Unscrambler s data file If the data file was originally imported from an NS
262. hances to influence the model PCA Dialog Validation Methods Validation is important to check whether your model will make a good fit on future data not used in the original computations A good model should generally describe data similar to that available when building the model Validation gives you an estimate of the error you can expect when fitting new similar data to the model It is also important when determining how many components are optimal for the data at hand In the Validation method field you choose between the 3 provided options as listed below Note Be cautious when you have a designed data table Do not use Cross Validation when there are few samples in the data table Validation methods alidation method Description Leverage Correction Quick and dirty Used for a rough initial estimate Do not use it for final validation as the leverage correction validation results are often too optimistic Cross Validation Simulates test set validation May be time consuming Allows you to use Martens Uncertainty test Test Set Validates the model against separate samples If you choose the method Cross Validation the Setup button next to it will be enabled Press it to enter the Cross Validation Setup dialog where you can further define your cross validation settings Tick in the box for Uncertainty Test only active if Cross Validation has been selected if you wish to take advantage of cross val
263. he Editor The area that is going to be overwritten is marked by a frame Supported Formats for Import The Unscrambler lets you import data from the following formats 50 e The Unscrambler Menu Options The Unscrambler Program Operation e Unscrambler Results see File Import Unscrambler Results p 51 e Unscrambler Data see File Import Unscrambler Data p 52 e Unscrambler5 Results see File Import U5 Results p 53 e Unscrambler5 Data see File Import U5 Data p 53 e Unscrambler ASCII files see File Import Unscrambler ASCII p 53 e Flat Flat wide ASCII files see File Import ASCII p 53 e Lotus files see File Import Lotus p 55 e Excel files see File Import Excel p 55 e Database see File Import Database p 56 e JCAMP DX files see File Import JCAMP DX p 59 e Tracker files see File Import Tracker p 61 e NSAS files see File Import NSAS p 61 e APC files see File Import APC p 61 e GRAMS files see File Import GRAMS p 61 e Matlab files see File Import Matlab p 61 e MVACDF see File Import MVACDF p 62 e Guided Wave files see File Import CLASS DA amp SpectrOn p 63 e Indico files see File Import Indico p 64 e Files from Brimrose instruments see File Import BFF3 p 64 e User Defined Import UDI see File Import User Defined Import UDD p
264. he Save Plot dialog Dialog Save Plot in use on a PLS regression result plot Save in amp Data E vl e c EJ E BugDesignTables a Tutorial Chapters 28Feb Ey wUTemp C drive BKUP Tutorial Dominique Sri dad7bf2256f Tutorials07 03 03 Ka E drive BKUP Tutorials Snapshots my Documents Gad Tutrials Final Program Files unscs8 0 2 Save as type Bitmap bmp z Cancel portable network grahics png TIFF Cat PC Op PC_02 4 6 PLS2 Regression Variable v Total PLS2 Regression Y var PC REDNESS 1 REDNESS 1 For Help press F1 4 Warnings Riw GU The Unscrambler Program Operation The File Menu e 49 File Import This option lets you import data from an external data file This may be a file from an earlier version of The Unscrambler or one with a different format e g Excel or ASCII Note The Unscrambler looks for files with the expected extension only You have to rename the file extension if the file you want to import has another How to Import Data When you have chosen which kind of file you want to import from the Import sub menu you enter the first dialog Import Target Once you have selected your import target and clicked OK you enter the Import dialog Import Target This dialog appears only if the currently active document is a non designed data table It gives you 3 choices regarding your import target e New data tab
265. he Unscrambler User Manual This tab opens the Table of Contents which is the default sheet Click your way through the main chapters Display more levels open a chapter by clicking on or close a chapter by clicking on Index ce re The Index tab lets you access the Index where you can search for alphabetically inde xed topics This may be useful for instance in the following cases 1 Direct access to a menu option which you want to lookup Type in the name of the menu e g file and under the word file in the list of topics you will find all the options that belong to the File menu 2 Locate a function whose name you are not sure of You may need to try several keywords before succeeding be imaginative do not give up Search The powerful search engine allows you to search for occurrences of one or several words Once you click on the List Topics button a list of all Help pages containing the selected word s is displayed as shown hereafter Finding topics containing the word constraint Contents Index Search Type in the keyword to find constraint Select Topic to display A Special Case Mixture Situations Analyzing Mixture Designs with PLS Constraints Between the Levels of Several Design Wizard Enter Multi Linear Constrai Designs based on a simplex D optimal designs D Optimal Designs for Optimization Purposes How To Deal with Small Proportions How
266. he set you are working with A new random selection of test samples is made each time a new analysis is run Uncertainty Test Dialog You can either decide how many Principal Components should be taken in consideration during the rotation or let The Unscrambler use the optimal number This selection is done by clicking on the button shown hereafter The Unscrambler Program Operation The Task Menu e 265 Dialog Uncertainty Test Uncertainty Test x Number of PCs to use in uncertainty test C Number of PCs to use aed Cancel How To Select Number of PCs to Use We suggest that you make a first model then remove obvious outliers and interpret results from the new model to find the optimal number of components to be used in the Uncertainty Test Make sure to look at individual Y variances in PLS2 situations The optimum number of PCs PLS components may be varying for the different Y Variables Individual PLS1 models often have better prediction ability than a global PLS2 model unless the Y variables are collinear Warning Limits Dialog During analysis a number of tests are applied to the results If the values calculated during these tests fall outside certain limits a warning is issued These limits are user definable i e you can change them by selecting Warning Limits in the appropriate Task dialog thus entering the Warning Limits dialog The default limits are to a certain extent heuristic and derived fr
267. heet This sheet lets you define various parameters depending on the type of plot Lookup the plot you are interested in among the sections below e Line Plot e 2D Scatter Plot e 3D Scatter Plot 134 e The Unscrambler Menu Options The Unscrambler Program Operation amo Software AS e Normal Probability Plot e Histogram e Matrix Plot e Special plot Percentiles e Special plot Mean and SDev e Special plot Predicted with Deviations Options Dialog Sample Grouping Sheet This sheet is available for sample oriented plots i e 2D or 3D scatter plots and normal probability plots of samples where the variables are model results Results Viewer or raw data plotted from the Editor Examples e 2D score plot in PCA e 3D scatter plot of raw data e N plot of Y residuals in PLS Sample grouping makes predefined groups of samples appear more clearly on the plot by means of colors or special symbols To start specifying your Sample Grouping options check the box Enable Sample Grouping at the top of the dialog box as shown below Dialog Options Sample Grouping sheet Options General Sample Grouping Iw Enable Sample Grouping OK Separate with Cancel C Colors e Special Symbols Group By C Calibration and Validation c C Design Sample Groups C Reference Samples Value of Variable X variables EN fi Number Of Groups 4 Markers Layout C Name TT MMMM DPT ys C Number C Symbo
268. hether you want to sort the samples by their names or the values of given variables and in the field Order whether the samples are to be sorted with the highest values at the top descending or the lowest value at the top ascending The Keys field is only enabled if you have chosen to sort by values Here you must specify the variable to use as 1 and 2 key The samples will first be sorted by the first key then by the second key Modify Sort Samples 3 D Sorting is disabled in 3 D tables with an OV layout since this would disturb the Primary Secondary structure of the samples Modify Sort Samples by Sets When using this option all samples belonging to Sample Set 1 are displayed in the first rows of the data table followed by samples from Sample Set 2 etc The sorting order is defined by the alphabetic order in which the Sets are displayed as a list in the Set Editor Samples that do not belong to any set are displayed at the end of the table If the sample sets currently defined in your data table are overlapping the command opens the Select Sample Sets To Sort dialog where you can choose which sets must be taken into account for the sorting operation Note1 This option is only available when at least one Sample Set is defined Note2 You cannot sort samples with locked cells e g designed data Undo Redo Modify Sort Samples by Sets with the corresponding options of the Modify menu Select Sample Sets
269. hree very small matrices they each only contain a vector of size A the number of principal components are included in order to maintain useful functionality in the file dialogs the variance plot the variance details dialog and the print details mechanism Preserving this functionality helps assessing the model quality when opening models for Prediction or Classification via file dialogs e The Micro model is the smallest possible it can only be used for Prediction To limit model size even more than with the Minimum model the results matrices are saved for the maximum number of PCs only i e PCmax This means that at the Prediction stage the number of PCs to use will be frozen to PCmax Note Restrictions on the Minimum and Micro Models The Minimum and Micro models do not contain enough result matrices to be opened in a Viewer Upon clicking OK in the Regression dialog a warning is displayed to remind you of the restrictions as shown below The Unscrambler j x You have selected a Minimum or Micro size for this J model This means you cannot open it in a Viewer the model can only be used for prediction and or classification Always create a Full or Compact model and verity it in the Inthe future do not show this message SS Cancel When the calibration is finished the View button is hidden making the Close button the only allowed action in the Progress dialog When clicking this button you are asked to save the mode
270. ialog is similar to the case of a factorial design see Add Design Variable Dialog Factorial or Plackett Burman Design Design Wizard Define Non design Variables Here you specify responses non controlled variables or variables with constant value The values of these variables are simply recorded during the experiments Click New to create a new variable Note Once the design is built you will still have the opportunity to create new non design variables in the Editor Edit Insert Variable or Edit Append Variables Add Non design Variable When you click respectively New or Properties in the Define Non design Variables dialog you enter a dialog called Add Non design Variable or Variable Properties respectively Here the non design variables are numbered automatically and you simply choose a name for the variable Define Non design Variables Next Dialog Once you have added all necessary non design variables by clicking New or changed their definitions by clicking Properties you may proceed by clicking Next You will then access the next dialog which varies depending on the type of design you are building e Full Factorial Plackett Burman Central Composite or Box Behnken Design Details e Fractional Factorial Design Type e D optimal Non mixture Mixture design Define Model Design Wizard Define Mixture Variables In this dialog you define the mixture variables that you want to investigate in your design
271. iations between the observed data values and the model approximation of those values An object sample or variable with large residuals is not well described by the model Residuals should be small and have no pattern The plots of Y residuals are only available for Regression Analysis of Effects and Response Surface results The Residuals dialog consists of up to three sheets General Outliers and Patterns The Patterns sheet is available for Regression results only The Outliers sheet is not available for PCA In the case of a Multivariate Curve Resolution MCR the two sheets are MCR Fitting and PCA Fitting 186 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Residuals Analysis of Effects Residuals Eg General Outliers Cancel Help CAMO rT Plot Y residuals vs Predicted Y m Y residuals Normal Probability Y residuals nenia Y residuals vs Score Influence Plot E Studentized Variance per Sample Samples Variable Residuals IZ Calibration Sample Residuals seen Sample and Variable Residuals T Validation r Sample EE La Y variable CY Cy Residuals Dialog General Sheet In the Plot field on the General sheet you must choose which type of plot you want to make by selecting the appropriate radio button You have 8 choices as described below Plots in the Residuals dialog General sheet Piot type Description
272. idation segments for assessing the stability of your results Martens Uncertainty Test Tick the button Hi access the Uncertainty Test dialog where you can set up the number of PCs to use in the Uncertainty Test this is not the same as the number of PCs computed in your analysis If you choose the validation method Test Set press Setup to enter the Test Set Validation Setup dialog where you can further define your test set validation settings PCA Dialog Miscellaneous Options In the bottom of the dialog you can find general options for your analysis 260 e The Unscrambler Menu Options The Unscrambler Program Operation nscrambler User Manual Camo Software AS 1 Model Size Number of PCs to be computed Center Data Add Start Noise Ao bo ba Issue Warnings Number of PCs In the Num PCs field you define the maximum number of Principal Components to compute Choosing many components will give you a model which explains more of the variations in the data but which also takes more time to build For a first computation we recommend that you choose a rather large number of PCs 1 2 or 1 3 of the total number of X variables for non spectral data around 10 for spectra so as not to miss the optimum in the variance curve Later on you may re calculate your model with fewer components e g one more than the optimal number so that the optimum still shows on the variance curve to limit the size of the model Model Siz
273. if you have selected a sample to plot or a Sample Set if you have selected a variable to plot from the drop down list in the Scope field If necessary press the Define button to define new sets in the Set Editor The data are plotted as a straight line through the point 0 50 if they are normally distributed Data points that are significantly different from the others deviate from the line and are placed in the upper right and lower left corner of the plot You may add such a line manually by using Edit Insert Draw Item Line Use Edit Options to change the layout of the plot Plot Histogram A This plot displays the distribution of the data points in a data vector as well as the normal distribution curve A histogram gives useful information when you are exploring raw data The height of each bar in the histogram shows the number of elements within the value limits of the bar The normal distribution curve is useful when you want to see whether the distribution of your data is skewed You may want to transform your variables if they are skewed in order to get a better model Select the sample or variable you want to plot and use Plot Histogram You enter the Histogram Plot dialog where you choose the scope of your plot Specify a Variable Set if you have selected a sample to plot or a Sample Set if you have selected a variable to plot from the drop down list in the Scope field If necessary press the Define button to
274. ification Dialog X variables Sheet Here you choose which Variable Set to use as X in your calculation in the Variable Set field You can either choose an existing Variable Set from the drop down list or define a new in the Set Editor which you enter by clicking Define Classification Dialog Pretreat Vars Sheet This sheet is useful if you want to perform automatic pretreatments at the classification stage Note that this option is only available if your class models have been calibrated on data pretreated in the Unscrambler Once you have selected relevant automatic pretreatments using the Pretreat button see Classification Dialog General Options specify which Variable Set to pretreat either by choosing an existing Variable Set from the drop down list or by defining a new in the Set Editor which you enter by clicking Define Note You are free to select a Variable Set to pretreat which differs from the X Variables set But make sure that it contains the same number of variables as the original set of variables pretreated prior to building the class models If not you will get an error message as shown below The Unscrambler Program Operation The Task Menu e 279 Dialog Invalid number of Pretreatment Variables The Unscrambler d Number of variables selected in Pretreatvar page should be equal to number of pretreatment variables in Model Classification Dialog General Options Tick the Centered Models box to make sure that
275. ilable if you have chosen the Uncertainty Test option in a multivariate regression model When using this function The Unscrambler can mark the significant X Variables in the model describing Y The marking of significant X variables is accessed either from the Edit menu Edit Mark Significant X Variables only or from the specific icon shown above This marking applies simultaneously to all relevant The Unscrambler Program Operation The Edit Menu e 147 plots on screen Loading X loading Weights and Y loading and finally the Regression Coefficients plot Marking is available for any kind of data For more general information about marking lookup section Edit Mark Basic Notions Beware of misinterpretations with PLS 2 In the case of PLS2 models significant X Variables may differ from one specific Y Variable to the other To avoid misinterpretation the marking is removed when the user plots a new Regression Coefficients plot in another window The un marking is also performed when plotting Regression Coefficients for another Y Variable Make sure to click the Mark Significant X Variables icon again while the Regression Coefficients plot you are studying is active Edit Mark With Rectangle When this option is active you can draw a rectangle in the plot All objects within the rectangle are marked For more general information about marking lookup section Edit Mark Basic Notions Edit Mark One By One E or
276. iled instructions Network Installation Multiple User Licenses Installing The Unscrambler on a network is a straightforward procedure in eight steps 1 oO So CE 9 Download of The Unscrambler License Server Installation of The Unscrambler License Server Request of an Unscrambler License Server Activation Key Activation of The Unscrambler License Server Definition of user names and passwords Configuration Settings Download of The Unscrambler Client Installation of The Unscrambler Client A detailed description of each of these steps is given below Downloading The Unscrambler License Server 1 Download The Unscrambler License Server from www camo com and save it into a Temp folder on your server Extract the file A folder The Unscrambler License Server is created Double click on file Setup exe to start the installation De Installation and Configuration The Unscrambler Program Operation he Unscrambler User Manual Camo Sof amp The Unscrambler License Server 9 5b Elte 2 File Edit view Favorites Tools Help Och H ki JO search E roses jas K Ka 1 E Address E O Camo The Unscrambler License Server 9 5b v Go Folders Name Size Type Date Modified a 0x0409 ini SKB Configuration Settings 25 02 2003 22 34 Datat cab 92KB WinZip File 27 10 2005 18 38 Ejlinstmsia exe 1 669KB Application 11 03 2002 20 15 Ejlinstmsiw exe 1780K6 Application 11 03 2002 10 06 l isscripts Msi 697 KB Wi
277. ill have to choose between starting from scratch modifying or extending an existing design If you chose File Duplicate As Modified Design the first dialog is Define Design Variables where you may add new design variables modify the properties of existing ones or remove them Design Wizard Dialogs The Design Wizard contains a dynamic sequence of dialogs where the type of dialog coming next often depends on your choices in the current dialog The best way to get help on a specific dialog is to click the Help button when using that dialog in The Unscrambler However if you wish you may lookup each individual dialog by selecting it from the list hereafter Alphabetical list of Design Wizard dialogs Dialog name Description on page Design Wizard Define Design Purpose p 112 Design Wizard Define Design Variables p 104 Design Wizard Define Mixture Variables p 107 Design Wizard Define Model p 111 Design Wizard Define Non design Variables p 107 Design Wizard Define Process Variables p 109 Design Wizard Design Details p 116 Design Wizard Design Type p 116 Design Wizard Design Type Mixture p Design Wizard Enter Multi Linear Constraints p 109 Design Wizard Extend a Centroid Design p 103 Design Wizard Extend a D optimal Mixture Design p 102 p 102 Design Wizard Extend a Fractional Factorial Design p 101 Design Wizard Extend a Full Facto
278. imal Mixture Design by adding reference samples center points or replicates Tick the radio button Add Center and Reference Samples or Replicates and press Next You will enter the next dialog Design Details Design Wizard Extend a Lattice Design In this dialog you have the choice between the following 4 extensions Add Levels on Category Process Variables This radio button is only available if you defined Category Process variables in your first design The next dialog will be Define Process Variables Add Center and Reference Samples or Replicates Center samples and replicates are added to increase the reliability of the model Add reference samples to see how your design samples compare The next dialog will be Design Details Increase Lattice Degree A Lattice design can be extended from degree 1 to 2 5 or from degree 2 to 4 The next dialog will be Design Type Mixture where you can set the desired degree Note Increasing the Lattice degree is only possible if your first design has a degree of 1 or 2 Higher degree lattices can only be extended to Centroid 102 e The Unscrambler Menu Options The Unscrambler Program Operation Extend to Centroid Design This choice first leads you through dialogs Define Non design Variables and Define Model Then you enter the dialog Design Type Mixture where the recommended extension design is Centroid Design with Interior Points Select this design and click Next The next
279. in different plots Option Relevant Plot Calibration Sample plots Variances Validation Sample plots Variances Previous Vertical PC PC dependent results Next Vertical PC PC dependent results Back to Suggested PC PC dependent results Next Horizontal PC PC dependent results Previous Horizontal PC PC dependent results X Variables Variable plots Variances X1 Variables Variable plots Variances Three way PLS results X2 Variables Variable plots Variances Three way PLS results Y Variables Variable plots Variances Residual Variance Variances 162 e The Unscrambler Menu Options The Unscrambler Program Operation Y Residuals Residuals Studentized Residuals Residuals View Source Calibration Pal Shows or hides the calibration samples This option can be combined with Source Validation to see whether the calibration and validation results are similar This option is only available in plots based on samples You cannot disable both calibration and validation samples View Source Validation Shows or hides the validation samples This option can be combined with Source Calibration to see whether the calibration and validation results are similar This option is only available in plots based on samples You cannot disable both calibration and validation samples View Source Previous Vertical PC 4 The command View Source Previous Vertical PC and i
280. in the Sample Name field an error message will be displayed prompting you for a correct entry see the Invalid Format dialog Fill in the names of your variables by double clicking on the upper margin where variable names are displayed See the figure below for an example of the ON layout O2V data table in the Editor STEI bal 562 000 910 449 3 EIEZ az 268 543 s2R1 2 89 10 00 ER Jl 5 90 IS3R1 Ia 3 55 6 90 IS3R2 lt z2 2144 247 iS4R1 4 39 1 83 54R2 Ia 2 88 6 33 File New 3 D data table variable and sample names The columns of a 3 D data table with layout OV are combinations of Primary and Secondary variables Thus Variable names consist of two parts referring resp to Primary Var and Secondary Var Name Similarly the rows of a 3 D data table with layout O V are combinations of Primary and Secondary samples Thus Sample names consist of two parts referring resp to Primary Sam and Secondary Sam Name The two parts of the name are separated by a hyphen Upon invalid entry in one of the two part name fields the Invalid Format dialog prompts you for a correctly built name as shown below The Invalid Format dialog Jack Foye This Value is Unacceptable The Valid naming format is Primary Field Secondary Field Primary Var Sam Name l Secondary Var Sam Name 285 1400 pe The Unscrambler Program Operation The File Menu e 47 File Open Ta el Citrl O
281. ing is disabled because it would change the dimensions of the data table Useful tips 1 Depending on your applications it may be relevant to average your data over all Secondary variables for an Ov layout resp samples for an O V layout To do so select the whole table then drag and drop it unto an empty spot of the workspace alternatively use File Duplicate As 2 D Data Table The resulting 2 D table can be averaged any way you wish If you reduce your data so that all Secondary variables are collapsed into an average for each Primary variable then you are left with Primary variables only and it is quite logical that the result should be stored as 2 D data table 2 You may use a similar trick to average in a flexible way over the samples ON layout or variables O V layout of your 3 D table The rows resp columns containing the averaged values can then be copied back to your original 3 D data table 232 e The Unscrambler Menu Options The Unscrambler Program Operation Modify Transform Transpose This command transposes the complete data table Designed data tables cannot be transposed Category variables are automatically split when you transpose a table containing such variables Note All defined sets are also transposed Transpose 3 D Transpose is disabled because it would change the dimensions of the data table Swapping the 3 D layout may be more relevant than a standard transposition Try it with Modify
282. ion i e the contents of a cell are likely to be influenced by its vertical neighbors In the Shift Variables dialog choose the appropriate variable set in the Scope field If necessary press Define to enter the Set Editor and define new sets 236 e The Unscrambler Menu Options The Unscrambler Program Operation Use the radio buttons in the Direction field to decide whether the defined set of variables is to be shifted up or down Specify by how many rows the variables should be moved in the Units field Note All rows of the table are affected by this operation The new cells created by this action are filled with m missing values as illustrated below Shifting variables creates new cells in the data table Before Variable Shift After shifting variable C down 1 A B C A B C 0 25 0 54 1 64 0 25 0 54 m 0 13 1 24 3 01 0 13 1 24 1 64 1 11 1 82 0 14 1 11 1 82 3 01 0 24 0 25 1 12 0 24 0 25 0 14 0 79 1 45 0 91 0 79 1 45 1 12 1 45 3 21 0 42 1 45 3 21 0 91 m m 0 42 Modify Shift Variables 3 D Shifting variables is disabled for 3 D data tables because it would change the dimensions of the data table Modify Reverse Sample Order This option reverses the order of samples in your data table from last to first All Sample Sets are automatically updated accordingly as with any other sorting option Note Using this option sorting is always pe
283. ions from the average this option is highly recommended You should have a good reason for choosing not to center your data before an analysis 272 e The Unscrambler Menu Options The Unscrambler Program Operation Add Start Noise You can select Add Start Noise if you want noise to be added to your data before the analysis is performed This option should only be used in special situations where a special data distribution makes it difficult for the algorithms to find a start direction Issue Warnings Finally The Unscrambler can give you warnings for several situations which may indicate poor validity of the results if you tick the Issue Warnings box For example the program can detect high leverages outliers in sample and variable space unusually low total explained variance etc The warnings are issued in accordance to the specifications in the Warning Limits dialog which you enter by clicking the Warning Limits button Note Remember that warnings are time and memory consuming Turn them off if you do not intend to use them Regression Dialog Miscellaneous Options MLR In the field Validation Method you can choose between the following validation methods 1 Leverage Correction 2 Full Cross Validation 3 Test Set Validation Exercise caution when you have a designed data table Do not use Full Cross Validation if your model is based on very few samples Since the only cross validation method available for MLR is Full C
284. ions which are mutually exclusive are selected via radio buttons Tick boxes are used to select multiple options For example you may center data and issue warnings at the same time 36 e General User Interface The Unscrambler Program Operation Plot Preview in Plot Dialogs Plot Preview Plot dialogs show you a preview of any plot type you are about to make like in the figure above enabling you to check that your choice of plot was correct You do however have to keep in mind that this is not a preview of your own data it just shows the general shape of the plot and indicates in which sub view the plot will be displayed Double clicking on the preview screen allows you to change the size of the plot window in the Viewer Extracted File Information Dialogs involving management of files on disk have an information field at the bottom The Unscrambler searches the files for information that tells you more about the contents of the file This information typically includes Type of file i f Preview screen for File name Size of the matrices true residual variance curve Directory name Set information Creator Weighting information Date of creation Calibration method Software version Validation method A preview screen see figure above in the information area of the file dialogs shows the true residual variance curve for the currently selected model whenever you select a result file that has the residual variance saved The do
285. is is that the minimum value is set as 0 and the rest are positive values To use this consistently for a set of samples make sure that the lowest point pertains to the same variable for all samples Linear Baseline Correction This transformation transforms a sloped baseline into a horizontal baseline The technique is to point out two variables which should define the new baseline These are both defined as 0 and the rest of the variables are transformed according to this with linear interpolation extrapolation You should take precautions not to select as basis variables that have spectroscopic bands As for the offset correction make sure that the lowest points pertain to the same variables for all samples Modify Transform SNV Standard Normal Variate SNV is a row oriented transformation which removes scatter effects from spectra by centering and scaling each individual spectrum Each value in a row of data is transformed according to the formula New value Old value mean Old row Stdev Old row In the dialog Standard Normal Variate select the Sample and Variable sets that define the matrix you want to correct in the Scope field The Define buttons take you to the Set Editor where you can define new sample and or variable sets Modify Transform Center and Scale Centering and or scaling your data may be useful to study the data in various plots from the Plot menu in the Editor or prior to running Task Statisti
286. ith a name or number Plot Header Draws the title of the plot and axis names In the field Plot Layout you define how the plot is drawn You select the type of plot you want by clicking on the appropriate radio button The available options are described below Note that not all radio buttons will be enabled in every plot Plot Layout options ption Description Curve Each object is plotted as a line Each data point is plotted as a bar Accumulated bars The bars from different objects are put on top of each other in the same variable Symbols Each data point is a symbol only Landscape The selected data is plotted as a landscape Contour The matrix is seen from above and the contours indicate value levels The matrix is seen from above using different colors to indicate different values 140 e The Unscrambler Menu Options The Unscrambler Program Operation In the Labels Layout field you choose how markers and labels will appear in the plot Available options are given below Labels layout options Description Select which part of the name to use in the plot The basis for the numbering is the current Set Variable Level The level of the sample is used to generate the name for example Indication Name This requires a designed data table Position on file The basis for the numbering is the raw data None No label Orientation field Horizontal or Vertical
287. ity Plot Source Bue Dous EI oe DK Cancel Matrix Tai 5 PCs 11 Samples Scores Help PCs Samples Note Only one vector at a time can be plotted as normal probability plot In the bottom right corner of the dialog a preview screen shows how the plot will look in the General View Note The preview plot is a general example and does not show what your actual data look like 216 e The Unscrambler Menu Options The Unscrambler Program Operation The Modify Menu Modify Compute General f Access this option from menu Modify Compute General or by clicking on the corresponding shortcut item in the Main Editor Toolbar This brings you to the Compute dialog which lets you perform arithmetic and more advanced computations on the whole data table or on selected samples or variables This option also helps in transforming variables Dialog Compute Compute Scope OK Samples SS Help All Samples 42 v Define Variables Dependent variables 3 e xpression Gel og Build Expression Examples V1 1 2 SIN V8 V9 1 S1 LN S2 X LN ABS X e In the Scope field select the Sample and Variable Sets to be affected by the computation Click Define if you need to define a new set this will take you to the Set Editor There are three ways of defining the mathematical expression to be applied You may 1 Type your expression directly in the Expression box 2
288. k Properties Dialog Provider Sheet In the Provider sheet of the Data Link Properties dialog select the database provider you want to import from 56 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog Data Link Properties Provider sheet E Data Link Properties Provider Connection Advanced All Select the data you want to connect to OLE DB Provider s IBM OLE DB Provider for DB2 MediaCatalogDB OLE DB Provider MediaCatalogMergedDB OLE DB Provider MediaCatalogwebDB OLE DB Provider Microsoft Jet 3 51 OLE DB Provider Microsoft Jet 4 0 OLE DB Provider Microsoft OLE DB Provider For Data Mining Services Microsoft OLE DB Provider for DTS Packages Microsoft OLE DB Provider for Indexing Service Microsoft OLE DB Provider for Internet Publishing Microsoft OLE DB Provider for ODBC Drivers Microsoft OLE DB Provider for OLAP Services Microsoft OLE DB Provider for Olap Services 8 0 Microsoft OLE DB Provider for Oracle Microsoft OLE DB Provider for Outlook Search Microsoft OLE DB Provider for SQL Server Microsoft OLE DB Simple Provider gt Next gt gt OK Cancel Help Hit Next to shift to the next dialog sheet Connection Data Link Properties Dialog Connection Sheet In the Connection sheet of the Data Link Properties dialog you locate the desired database from the proper server and specify the security settings for logging on
289. l Variable Level Indication Name e Value Of Variable sey ariables vii Select Position on File The Unscrambler Program Operation The Edit Menu e 135 In the field Separate with you can choose to visually differentiate between groups of samples either with colors or with symbols The Special Symbols option is especially useful when printing out plots in black and white The Group By field lets you define how the plotted points are to be divided into groups The available options are presented in the table below Group By options for Sample Grouping Options Description Calibration and Validation Uses different symbols colors to separate between calibration samples validation samples and samples used for both calibration and validation Cross Validation Segments Uses different symbols colors on each cross validation segment Design Sample Groups The groups correspond to the levels of one or several chosen design variables Reference Samples Uses different symbols colors to separate between the reference samples and the other samples Value of Variable Samples with the same level in this variable get the same symbol color For continuous variables up to 5 groups can be created If you take the option Design Sample Groups the Groups button is activated Click this to access the Sample Subgrouping dialog where you can select which design variable s must be used to generate the
290. l Primary variables Display Variable statistics if you want a summary of each Primary variable across all Secondary variables View Variable Statistics Displays a few descriptive statistics about your variables in a new Editor You first access the Variable Statistics dialog where you must select a Sample set The computed statistics are e Number of missing values e Minimum e Maximum e Mean Standard deviation e Skewness The results are displayed in a new Editor This new Editor behaves as a slave of the source data Editor As a consequence columns marked in the original data table will also be marked in the Variable Statistics table and vice versa An example of Variable Statistics results Red Blue Comp a ofMissing a 0 0000 0 0000 Min 2 0 0000 0 0000 0 0000 Ma 3 28 0000 44 0000 20 0000 Mean Lal 15 0000 19 3200 11 4286 isDev 5 7 4833 17 1786 7 4801 Skewness 0 2663 0 2050 0 1567 The usual File Edit View Modify and Task operations are not allowed on this slave Editor only Plot is possible Besides closing the Editor containing the source data table will also close the Variable Statistics table If you want to manipulate these statistics further you have to select rows or columns of the Variable Statistics table and drag n drop them onto an empty spot of the workplace This will generate a new table which you can freely manipulate edit and save View
291. l View 211 lot 2D Scatter General NIEWT ee egeesedh eege E E E EEN 212 Plot 3D Scatter General View 213 Plot Histogram General View secsccssssscescsssscveciscsssteevecesssacessaeves cocosenvn a anveveseiassivesisoseoses 214 Plot Matrix General View 215 Plot Normal Probability General View 215 d Rue e E EE 217 Modify Compute General 217 Modify Et E 219 Modify Transform Smoothing 0 0 cece ceeeeenee ceeeeaeeseeeeeaeeseeeesaeeseseeeesseseceesseeaeeeesagens 220 Modify Transform Smoothing Moving Average 0 ee seeceeeeseesceseeeeeeeeceeetseeeaeeasens 220 Modify Transform Smoothing Savitzky Golay 0 0 ce ce cceeseseceeeeeeeeseeneeeeeaeeneeeneeees 221 Modify Transform Smoothing Median Filter ee eee ce ceeseeeeteeceeeeeeaesneeeeeaees 221 Modify Transform Smoothing Gaussian Filter cece cee cneeeeeeeeseeeeseeeseeeeeees 222 Modify Transform Normalize cccecceseeseescceeceeesesecceeceeeeeseecseseaeeaeees caaeeaeceeeeeeceeeeneeaas 222 Modify Transform Spectroscopic Transtormaton cece sees ceseeeeseeeceeeseeeeeaeeseens 224 Modify Transform MSC EMSC uu ececcececeseceseeses caeceeeaeeseseesaeeaeeeecaesaeesnessassecesaeeneeaes 225 Modify Transform NOISE oo cece csceeceseee saeeeceseaeeeesceeeceesseceessnesaeeeeeeceseesaeseseeeaeenes 228 Modify Transform Derivatives ce eeceeeeescees seeeeeseceeeeeeseceesesseeeeeae seeeeesseseeseneaeenes 228 Modify Transform Deriva
292. l as usual Later on if you try to open the model from the File Open or Results dialog a message will pop up informing you that the model cannot be viewed The limited number of result matrices in models of reduced size also has some effect on the Export Model functionality For Minimum and Micro models this option is only available by selecting a model in the Results dialog and clicking the Export button For a Minimum PCA model only the Tracker format is supported For a Minimum regression model ASCII MOD and Tracker are supported For a Micro regression model only ASCH MOD is supported Minimum and Micro models support automatic pre treatment at the Prediction or Classification Minimum only stage Use the Results dialog and click the Properties button to register the transformations for automatic pre treatment Warning You should select model size Micro only after having computed and validated a similar model with size Full or Compact Use the optimal number of PCs determined from that previous model as Num PCs total number of PCs to be computed for the Micro model This way you will ensure use of the correct number of PCs at the Prediction stage Center Data Ticking the Center Data box makes a model where the data have been centered first i e the average of each variable is subtracted from each of the data values in the variable Centering means that the results can be interpreted in terms of deviat
293. l screen 294 go to 293 identification 295 shift to other window 1 2 297 tile horizontally 294 tile vertically 294 warning list 295 Windows 2000 29 Windows 95 29 Windows NT 29 Windows XP 29 wizards 37 workplace 29 X X and Y variance 179 X or Y variance 178 X or Y variance three way PLS 179 X Y variable numbering 154 X1l and X2 variance 179 X Y relation outliers 193 X Y Relation Outliers dialog 193 Z zoom in 156 zoom out 156 The Unscrambler Program Operation Index e 329
294. lat Wide ASCII import if you have ASCII files you want to import from other programs File Import ASCII You may import data from ASCII files with different formats provided that you specify as much as you know about the nature of the ASCII file you want to import This is done in the field File format in the dialog Import ASCII see the figure below which you enter when you click Import in the Import dialog An ASCII file can be flat or flat wide Note Files with extension csv can now be imported using File Import ASCII Flat ASCII Stream files are simply strings of numbers with no delimiters It is therefore vital that you know the exact size of the matrix The Unscrambler reads the first sample with as many numbers i e variables as you have specified in the variables field then proceeds to the next sample These files cannot contain names only numbers The Unscrambler Program Operation The File Menu e 53 Flat Wide ASCII Block files use a paragraph character to denote the end of each sample The file may also have header lines giving information about its contents You must enter the number of header lines or they may be imported into the matrix as data Dialog Import ASCII Import ASCII x File format OK Flat ASCII stream oS No of samples fi Cancel No of variables fi Help CW 0 Header lines to skip 10 IV First row is variable names IV First column is sample nam
295. lations from the Sample Set drop down list Statistics are calculated for the entire chosen sample set If you want to define a new set click Define to launch the Set Editor dialog If you want to exclude some of the samples from the analysis type in their numbers in the field Keep Out of Calculation or click Select to choose them directly from an Editor the Select Samples dialog Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10 13 27 Note How to use Sample Set and Keep Out e In Sample Set you select the general sample population the analysis applies to e InKeep Out you remove one or more of these samples if it turns out that they disturb the results The Unscrambler Program Operation The Task Menu e 253 The computations will always be performed on a default group of samples containing all the samples you have included in the analysis This group is called All samples In addition they can also be performed on smaller groups of samples determined by the levels of one or more of the variables in your data table The variables which have the ability to generate groups of samples are leveled variables category variables and design variables For instance if you are computing statistics for variable Quality and your data table contains a category variable called Equipment with two levels New and Old your statistics result file will contai
296. le e Current data table from cursor position e Current data table from origo Make your choice and click OK The imported data is inserted as the leftmost variables or the upper samples when you choose to insert them from origo Note Importing data into the current Editor will overwrite any cells that are covered by the imported data area If you do not want to overwrite these cells use the Insert or Append commands to insert new samples variables before you begin importing data Import Dialog Once you have selected your import target and clicked OK you enter the Import dialog which is identical to the Open File dialog except that the buttons Print and Lab Report are not available Select the file which contains the data that you want to import and press Import Dialogs differing according to the type of file allow you to select which matrices to import and give any other information The Unscrambler needs to import the data Import With Drag n Drop Instead of going via the File Import menu you can import data by using Drag n drop You simply mark data in another Windows program and drag it into The Unscrambler s Workspace or into an existing Editor document If you drop the data table onto the Workspace an Editor will be opened automatically You select whether you want to insert the data as columns or rows The columns or rows are appended at the end of the existing data table You may also overwrite the existing data in t
297. le Properties The Constraints Sheet Dopt test 3vars Properties xi General Notes Log Instrument Constraints This sheet summarizes the constraints associated to your design variables Multi linear constraints restrict the joint variations of two or more variables In addition mixture variables are linked by the mixture constraint which is that the sum of all mixture components is fixed equal to MixSum Multi linear constraints 1 000 T emperature 3 000 Time 3 000e 04 gt 0 MixSum fi 00 00 Mixture design Cancel Apply Help Select Transformations Dialog This dialog is accessed from the Pretreat button in the Classification or Prediction dialog as well as the Select button in the Transformations sheet of the Properties dialog It allows you to select the transformations that you wish to apply to the data as automatic pre treatments upon classification or prediction You may select among the available transformations either manually or with the help of the Select All button Clicking Select None tells the system not to use any of the transformations Your choice is validated once you click OK File System Setup This option lets you enter the System Setup dialog as seen in the figure below where you can configure The Unscrambler to suit your own preferences when you work Any changes you make will affect only your own user account and all files you work on Pressing Apply brings th
298. le matrix is selected Look at the cursor shape or the rows columns numbers to see whether you are in row or column mode Sample names will also be selected if you are in sample mode and vice versa Edit Adjust Width Hl Ctrl W Adjusts the width of selected columns to exactly fit all data values The Unscrambler Program Operation The Edit Menu e 133 If you want this command to also take variable names into account when adjusting the column width this must be specified in the System Setup dialog Edit Add Plot or Ctrl A In the General Viewer Edit Add Plot allows you to add a second plot of the same type as the one s already displayed This command is also available through the shortcut CTRL A Note This command can only be applied to plots of the following types Line 2D scatter and 3D scatter Edit Options fel Ctrl L You can use this command to change the appearance of the plot you are working on The Options dialog always contains the sheet General in addition the sheet Sample Grouping is available in predefined plot mode when you display a sample oriented plot e g a score plot Dialog Options Options xi General Sample Grouping M Grid T Equal Scale Cancel I Curve ID Help Vertical Line See T Connection Line M Plot Header r Markers Layout CG Name MIMI TTTTTTTITT Number Symbol Variable Level Indication Name Position on File Options Dialog General S
299. le name Note 1 The unit conversion option is disabled if the selected variable set has names all or strings Note 2 Unit conversion is not supported as an automatic pre treatment If some of the variable names cannot be interpreted as wavenumbers or wavelengths the following error message is displayed when attempting Unit Conversion The Unscrambler E xj i One or more variable names were not recognized as possible J wavenumbers and wavelengths Unit Conversion applies on numerical values for wavenumbers or wavelengths and the conversion is not performed Modify Transform MSC EMSC MSCorrection is a transformation method used to compensate for multiplicative and or additive scatter effects in the data EMSCorrection is an extension to regular MSCorrection using prior knowledge that includes extra parameters that can account for the physical or chemical phenomena that affected the spectra Both methods are row oriented transformations that is to say the contents of a cell are likely to be influenced by its horizontal neighbors Dialog Multiplicative Scatter Correction Multiplicative Scatter Correction Scope Samples All Samples Define Cancel Variables Help Define C Use existing MSC or EMSC model Compute and use new MSC model Msc Function Full MSC X xX a b C Common offset X X a C Common amplification X X b I Enable EMSC Test samples Omit importan
300. lection dialog On both sheets you also need to define the Weights for your variables in the Set Weights dialog The weighting determines the role of each of the variables you can change their influence on the model Consider weighting 1 SDev if you want all variables to have equal chances to influence the model Note You cannot have the same variable in both the X variable and the Y variable set Regression Dialog X Variables Sheet and Y Variables Sheet MLR These sheets are essentially identical to the sheets for PLS1 PLS2 and PCR but for MLR the weights are set to All 1 0 by default This means that the Weights button is disabled Note You cannot have the same variable in both the X variable and the Y variable set Regression Dialog Validation Methods PLS and PCR Validation is important to check whether your model will make a good fit on future data not used in the original computations A good model should generally describe data similar to that available when building the model Validation gives you an estimate of the error you can expect when fitting new similar data to the model It is also important when determining how many components are optimal for the data at hand In the Validation method field you choose between the 3 provided options as listed below Note Be cautious when you have a designed data table Do not use Cross Validation when there are few samples in the data table 270 e The Unscrambler Menu O
301. ler A 1 Installing The Unscrambler Geste eco eti sri E a REE EE E EE EE H Requesting an Unscrambler Activation Ren 5 Activating The Unserambler ccc eceeceeee seceeceeeseseeseeseeseeaecaeeseeesaeeeceesaesesesesaeeeeaesaee cues 7 Defining User Name and Dassword 8 Network Installation Multiple User Licenses ceeesesssssecesececeeeeseceeeeeesaeeseceseceeeeeeeeaeeeeeeeeeateeas 8 Downloading The Unscrambler License Server 8 Installing The Unscrambler License Server 9 Requesting an Unscrambler License Server Activation Ken 12 Activating The Unscrambler License Server 13 Defining User Names and Passwords cccecesessesseseeseseeeeceesaceeeseesacneesecesaeeeeaeeeeeesaseaeeaes 14 Configuring SECM Eeer eeng Eege Eed 15 Downloading The Unscrambler Cent 15 Installing The Unscrambler Client 16 Troubleshooting under Installation sssrinin rre ar NE E OE EEE E ri E 19 Symptom 1 InstallShield Error Message oo cceceesessceeeeceeseeeeeseeseeeeesesseeersaeeeesceeseeeaeeees 19 Symptom 2 Invalid Activation Key ccc sssscsscsesssssseeseesssesessssscceesseesseses REENEN 20 Symptom 3 Incorrect License Details in Standalone Installation eee serrer 21 Symptom 4 The Unscrambler Previous Version was not Upgraded to the New Version 21 Symptom 5 No Activation Key is Required in Standalone Installation eee 21 Upgrade from Trial Installation to Permanent License 0 0 cece ceeseseceeeseseeeeeeeeeeeceeseaeeecneeeeeeeenaes 22 System
302. les and variables to be transformed If necessary press Define to enter the Set Editor and define new sets In the Parameters field choose the Derivative order i e whether you want the Ist 2nd 3rd or the 4th derivative of the samples from the drop down list Then select the required Gap size and Segment size Note e The segment size must be an odd number for 2nd or 4th derivative e The gap size must be an odd number for Ist or 3rd derivative Modify Transform Derivatives Savitzky Golay Savitzky Golay differentiation computes derivatives of up to the AN order based on a polynomial approximation of a portion of the curve Dialog Savitzky Golay Derivatives x i Samples All Samples D Define __ Cancel Variables Help Spectra 100 Define Parameters Derivative order 1st derivative x Smoothing points Number of smoothing points 4 Number of left side points fil Number of right side points P The number of smoothing points is 4 Polynomial order P 4 The Unscrambler Program Operation The Modify Menu e 229 Make the appropriate choices in the Savitzky Golay Derivatives dialog by first selecting the sample and variable sets that define the matrix you want to differentiate in the Scope field If necessary press Define to enter the Set Editor and define new sets In the Parameters field choose the Derivative order i e whether you want the I
303. les by sets 237 swap 3 D layout 239 swap samples amp variables 239 toggle 3 D layouts 241 transform 219 undo 242 modify existing design 98 Modify Primary Variable Set dialog 250 Modify Sample Set dialog 250 Modify Secondary Variable Set dialog 250 modify supervisor dialog 25 Modify Supervisor dialog 25 modify user dialog 25 Modify User dialog 25 Modify Variable Set dialog 250 most recently used file list 91 move data directories 24 moving avaerage segment 311 moving average 220 322 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual Moving Average Smoothing dialog 220 MSC 225 MSC dialog 225 226 MSCorrection 225 multi linear constraint 307 Multi Linear Constraint 107 108 307 multi linear constraints 89 multiple linear regression 268 270 multiplicative scatter correction 225 multivariate curve resolution 267 Multivariate Curve Resolution dialog 267 multi way 307 MVACDF files export 74 import 62 N name qualifier 53 54 name variable 105 network installation 8 new data file 45 design 97 samples 125 variables 125 New Data Table dialog 45 new file 45 New Primary Variable Set dialog 249 New Sample Set dialog 247 New Secondary Variable Set dialog 249 new user dialog 25 New User dialog 25 New Variable Set dialog 248 next horizontal PC 168 next vertical PC 164 noise 228 260 262 271 273 309 Noise dialog 228 non designed data table restrictions 110 Non de
304. lex for instance if you defined Multi Linear constraints a D Optimal design will be generated The process is the same as in the case of a D Optimal Non mixture design see Generate Design D optimal Non mixture The experimental region is a simplex and the design includes Process variables In this situation The Unscrambler offers you the choice between e combining all points of the simplex mixture design with all combinations of levels of the process variables tick the radio button Full Factorial Combination e reducing the number of experiments by using the D optimal algorithm tick the radio button D Optimal Design Wizard Generate Design Mixture situation Design Wizard Generate Design x Choose if you want the mixture part of the design to be combined with all levels of the process variables or if you prefer to reduce the number of experiments with a D optimal algorithm Full Factorial Combination Design Points 1 Replicates D Optimal fat a Design Points 1 Replicates Number of Center Points 3 Above you can choose the number of points for your design Then press the Generate button to start generating D optimal design Condition Number to correctly explore the experimental region 1 32e 03 Choosing a solution with a lower condition number will result in reducing the explored region Select a Design to Use 31 Design Points Condition No 1 31e 03 30 Design Points Condition No 1 32
305. lication 11 03 2002 20 15 Slinstmsiw exe 1780 KB Application 11 03 2002 10 06 l155cript8 msi 697 KB Windows Installer P 09 04 2003 23 33 220KB Application 27 10 2005 19 02 Setup ini ZER Configuration Settings 27 10 2005 19 02 CH The Unscrambler msi 1 042KB Windows Installer P 27 10 2005 19 02 If you have received The Unscrambler installer on CD ROM place the CD in your CD player 4 An InstallShield Wizard dialog opens InstallShield Wizard The Unserambler Setup is preparing the InstallShield Wizard which EI will quide you through the program setup process Please wait a Configuring Windows Installer d BRRRRRRRRRRRRRRRR Cancel 5 The InstallShield Wizard for The Unscrambler is launched 2 e Installation and Configuration The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS InstallShield Wizard ES Welcome to the InstallShield Wizard for The e Unscrambler The InstallShield Wizard will install The Unscrambler on your computer To continue click Next 6 Read the licence agreement carefully and click Yes to accept the terms of the agreement InstallShield Wizard License Agreement Please read the following license agreement carefully Press the PAGE DOWN key to see the rest of the agreement MPORTANT READ CAREFULLY The CAMO license agreement is a legal agreement between you the Licensee and CAMO PROCESS AS the Licenser and the company you bought the s
306. ll Plots if you wish to print out all plots currently shown on screen The Print dialog for plots Print lz EI E Properties Name Lexmark NT Status Default printer Ready Type Lexmark Optra plus PS2 PC1 LexMark Where NWOSLO Comment NWOSLO Q_LEXOSLO Copies m Print j C Current plot Number of copies 1 All plots J Cancel Select the printer you want to use from the drop down printer list The properties of the printer can be viewed by pressing Properties See your Windows documentation or printer manual for information on setting up your printer You can print to a file by clicking on the Print to file box File Print Preview Lal It is a good idea to preview a document before sending it to the printer Print preview lets you take a look at how your pages will look when they have been printed The option is only available if a file is currently open File Print Setup In the top field of this dialog Printer you specify which printer to use for your print out Press the Properties button to enter a dialog where you may specify details about the print out such as orientation multipage printing etc In the Paper field you specify the size of the paper and the paper source if necessary Finally specify whether to print out as portrait or landscape by selecting the appropriate radio button in the Orientation field File Lock This command disables all changes t
307. ll conditioned and it may be difficult to estimate the model parameters The model does not describe the data well Some variables are badly fitted Variable Outlier Limit Calibration Variable Outlier Limit Validation Some variables are badly fitted 266 e The Unscrambler Menu Options The Unscrambler Program Operation he Unscrambler User Manual Camo Software AS Task MCR In the Multivariate Curve Resolution MCR dialog you define the data on which the model will be based and specify the technical features of the computations The dialog contains two sheets Samples and Variables and two general groups of options Enable Initial Guess and Constraints Setting Dialog Multivariate Curve Resolution MCR Multivariate Curve Resolution MCR d x Samples Variables OK ok Sample Set Cancel Gees e 36 Define Keep Out of Calculation Select Initial Guess Samples T Enable Initial Guess Concentrations Variables Wel m Constraints Setting IV Non negative Concentrations Variables IV Non negative Spectra Samples I Closure I Unimodality Sensitivity to pure components 100 MCR Dialog Samples Sheet In the Sample Set field select the sample set on which to perform the calculations Use Keep Out of Calculation if you wish to leave out some of the samples If you have enabled an initial guess based on some known spectra
308. log 284 Prediction dialog 209 282 283 preview screens 37 previous horizontal PC 167 previous vertical PC 163 Primary Sample 309 Primary Variable 309 principal component 309 principal component analysis 258 Principal Component Analysis dialog 258 principal component regression 268 269 print 82 lab report 82 preview 83 range 82 selection 82 setup 83 Print dialog 82 Print Lab Report dialog 82 Print Topics dialog 300 process variable 309 Progress dialog 254 project system setup 90 projected 158 properties 84 244 constraints sheet 89 general sheet 84 instrument sheet 87 log sheet 86 model sheet 88 notes sheet 85 X Y relation outliers 193 sample 244 plots transformations sheet 86 87 ID 34 variable 244 324 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual Properties dialog 84 244 proportional noise 309 Q Quit The Unscrambler 96 R radio button D Optimal 114 115 Full Factorial Combination 113 114 115 random order 309 randomization 117 309 Randomization Details dialog 117 range normalization 222 223 raw data 171 recalculate with marked 287 with passified marked 288 with passified unmarked 288 without marked 288 recent files list 95 redo 243 reduce average 233 Reduce Average dialog 233 reduce result file size 91 Reducing number of experiments 114 115 reference points 116 117 reference sample 309 reference samples 116 117 use in significance tes
309. log 289 reverse marking 148 reverse sample order 238 reverse variable order 239 RMSE 180 RMSEC 310 RMSED 310 RMSEP 310 plots 178 rotate 157 The Unscrambler Program Operation Index e 325 S sample 310 names 153 outliers 181 statistics 151 sample grouping sheet 135 Sample Outliers dialog 181 Sample Properties dialog 244 sample range error 262 Sample Statistics dialog 153 save 49 save plot 49 Savitzky Golay Derivatives dialog 230 231 Savitzky Golay 221 Savitzky Golay Smoothing dialog 221 scaling 155 310 314 exact frame 156 exact min max 156 frame 155 min max 155 transformation 232 Scaling dialog 155 scores 183 2 x 2D scatter plot 184 4 x 2D scatter plot 184 general plot 184 scores and loading weights 182 bi plots 183 four plots 183 two plots 183 Scores and Loading Weights dialog 182 scores and loadings 181 bi plots 182 four plots 182 two plots 181 Scores and Loadings dialog 181 Scores dialog 183 screen layout 31 Secondary Sample 311 Secondary Variable 311 segment 311 segments 160 select all 133 bars 149 cell 31 continuous range 31 non continuous range 31 range of cells 31 samples 31 132 variables 31 133 Select Design to Extend dialog 100 Select Design to Modify dialog 100 Select Design Type dialog 98 Select Existing Design dialog 100 Select Method to Use dialog 97 98 select sample sets to sort dialog 237 Select Sample Sets to Sort dialog 237 Select Samples dialog 132 252 Select Sets dialog
310. ls Note1 You cannot extend a 3 variable Axial design with End Points to a 3 variable Centroid Design with Interior points these two designs are strictly identical Note2 It is not possible to create a Centroid design if you have more than 6 mixture variables The Unscrambler Program Operation The Design Wizard e 103 Design Wizard Define Design Variables In this dialog you specify the design variables that you need to have in your new design Dialog Define Design Variables Design Wizard Define Design Variables You are now ready to specify your design controlled variables Click New to enter anew variable Click Properties to modify an existing variable ID Name Data Type Levels A Stirring Speed Continuous 20020 D Delete Properties Information Fractional Factorial Design 1 Design Variables ig Cancel Help The top field displays information about all defined design variables while the Information field displays information about the currently selected variables and design The dialog allows you to add a new or to delete or change an already defined variable by clicking one of the buttons New Delete or Properties Clicking New or Properties leads you to the Add Design Variable or Variable Properties dialog where you can specify the name and properties of the design variable For Central Composite designs this dialog also includes an extra field star points distance
311. ly If you select an existing set and press the Properties button you can change the set definition in the Modify Sample Set Modify Variable Set dialog This is identical to the New Set dialog 244 e The Unscrambler Menu Options The Unscrambler Program Operation amo Software AS Unscrambiler User Manual Modify Edit Set 3 D data table Since 3 D data tables have an extra dimension the Set Editor dialog allows you to edit three types of user defined sets instead of two For an OV table these are Primary Variable Sets Sample Sets and Secondary Variable Sets as shown below Dialog Set Editor 3 D data table Set Editor There are two categories of sets for each vector orientation samples variables The predefined sets are maintained automatically by the program and can not be modified or removed by the user The userdetined sets however can be freely manipulated Show sets of type Userdefined E Primary Variable Sets EN Primary Variable Sets Data Type Sample Sets Secondary Variable Sets Add em Cancel Help Note Since no analysis is allowed on a 3 D table with layout O V the Set Editor is disabled for such tables New Sample Set Dialog Allows you to specify a new sample set or modify its properties In the General section give the new set a name in the Name field This is the name that will refer to the set in other dialogs when you want to make models plot data etc It should therefo
312. ly interactions free of counfounding can be plotted Plot Statistics The Statistics dialog gives you access to all plots available after you have run a Statistics calculation Select the tabbed sheet that gives you the information you are looking for The available sheets are 1 General 2 Compressed 3 Group only available for designed data 4 Scatter only relevant for spectra The Unscrambler Program Operation The Plot Menu e 205 5 Cross Correlation only available if you ticked the Calculate Cross Correlation box in the Statistics dialog from Task Statistics Statistics Dialog General Sheet On this sheet you can specify individual plots of the types listed below either for all samples or for pre defined groups or subgroups of samples In the Plot field you can choose between different plots as displayed below Individual plot types Plot type Description A plot of the average of each of the variables tandard Deviation A plot of the standard deviation of each of the variables Mean and Std Deviation The mean is plotted as bars with the standard deviation indicated by lines on top of he bars Shows the percentiles as a box plot A plot of the standard deviation of replicated samples is displayed Precision Next you choose whether you want to plot values for the different subgroups adjacent to each other Line Plot or generate a new dimension in the plot
313. matically be kept out to make the dataset fit into the selected model Prediction Three Way PLS Dialog Sec X vars Sheet Choose which Variable Set to use as Secondary X for prediction in the Variable Set field To define new variable sets click Define to go to the Set Editor If some of the Secondary X variables were kept out from the model at the calibration stage they will automatically be kept out at the prediction stage and their numbers will be displayed in the grayed out field The Unscrambler Program Operation The Task Menu e 285 Keep Out of Calculation This occurs when you select the model to use see Prediction Three Way PLS Dialog Model The following warning will then be displayed The Unscrambler L Some of the Sec X variables will automatically be kept out to make the dataset fit into the selected model Prediction Three Way PLS Dialog Y reference Sheet Tick Include Y Reference on this sheet if you have reference measurements that you want to use in the prediction This will activate the Variable Set field where you specify which set to use as Y reference either by choosing an existing Variable Set from the drop down list or by defining a new in the Set Editor which you enter by clicking Define These reference values are used only as a comparison to assess the quality of the predictions not to guide the predictions Prediction Three Way PLS Dialog Model You select the model to use in the predi
314. mber of data points variables whereas NSAS Vision assumes these parameters to be expressed in wavelengths nanometers Therefore you must multiply The Unscrambler s parameter with the difference in wavelength between two neighboring wavelengths in the original data table before you put it into the NSAS Vision parameter field An exception to this is NSAS Vision Savitsky amp Golay smoothing which assumes number of data points as segment size just like The Unscrambler Vision Savitsky amp Golay 1st and 2nd derivative assumes the segment size parameter to be expressed in number of data points for the smoothing part and in nanometers for the derivation part of the transformation NSAS Vision transformations and corresponding Unscrambler transformations Wees Math NSAS Vision Math Unscrambler Transform Name Comments ype ype Name 2 Reflective energy Spectroscopic Absorbance to NSAS multiplies result by Reflectance 100 to express it in Compute Multiply by 100 0 3 Kubelka Munk Smoothing Moving average Spectroscopic Absorbance to Reflectance Spectroscopic Reflectance to Kubelka Munk 4 1 derivative No equivalent NSAS Vision derivation disregards distance between points or segment centers ie derivative results are unscaled The Unscrambler s Norris Derivation gives similar results but applies a slightly different smoothing and scales divides by number of points between segment centers
315. mmand You cannot find or replace values located in the row or column headers Sample names or Variable names Edit Correct Mixture Components This function is specific to designed tables It is available only when one or more mixture variables are selected This function corrects the selected mixture variables so that the sum per row i e per experiment of all Mixture variables is equal to the MixSum value The values in the non selected mixture variables remain fixed during the operation Edit Convert to Category Variable In the Editor highlight one or several columns containing continuous variables and choose Edit Convert to Category Variable This launches the Convert to Category Variable dialog where you can select a relevant conversion principle Dialog Convert to Category Variable s Convert to Category Yariable s Note This option is disabled for 3 D data tables with an OV layout Convert Individual Values into Levels If you choose the first option the values of the source variable will be sorted in increasing order and each value will be assigned a level index accordingly Visually the only obvious change in the Editor is the blue variable name and the left justified cell values Warning If the source variable has more than 20 different values the remaining values after sorting and affecting level TS indices 0 19 will be replaced by m missing Convert Ranges of Values into Levels
316. mples on other plots Another is to mark ranges of the spectra in the Important variables plot to make a new model based on only important wavelengths Note If your Viewer contains more than one plot marking is only possible from the currently active sub view For instance if the currently active sub view contains a score plot you have to click on the sub view containing a variable plot before you can mark any variables Once you have marked objects they appear marked in all current and future plots until you unmark them or close the Viewer Edit Mark Outliers Only All outliers found by The Unscrambler s automatic outlier detection are marked The limits used for detecting the outliers are set in the Warning Limits dialog You access this dialog from all model dialogs Edit Mark Test Samples Only 3 or Ctrl T All test samples are marked The option requires that the model used to create the plot is made using test set validation For more general information about marking lookup section Edit Mark Basic Notions 146 e The Unscrambler Menu Options The Unscrambler Program Operation Edit Mark Evenly Distributed Samples Only This command lets you mark a representative subset of the samples in any plot of samples You specify which samples to mark in the dialog Evenly Distributed Samples which is displayed below Dialog Evenly Distributed Samples Evenly Distributed Samples Method z
317. mport Target dialog 50 Import Worksheet dialog 55 important variables 195 Important Variables dialog 195 Important Variables dialog three way PLS 195 Indico files 64 Information dialog 122 initial guess 267 input range 262 insert 122 category variable 124 line 144 mixture variables 125 sample 123 text 144 variable 123 installation network 8 standalone 1 instrument parameters 76 79 interaction and square effects 277 Interaction and Square Effects Selection dialog 111 251 Interactions and squares 111 intercept 305 interior point 305 issue warnings 260 262 271 273 276 item delimiters 53 54 item properties 144 Item Properties dialog 144 J JCAMP DX Import dialog 59 JCAMP DX files export 72 import 59 import 3 D 69 K kill subwindows on close 91 K means clustering 255 L large plots 92 93 Last Checks dialog 118 lattice design extend 102 layout 245 Layout dialog 245 leveled variables 306 levels 306 leverage 190 correction 260 leverage correction 306 Leverage dialog 190 license change 301 licensee 302 Licensee Information dialog 302 limits for outlier warnings 306 line plot 175 general view 211 Line Plot dialog 175 Line Properties dialog 144 The Unscrambler Program Operation Index e 321 loading weights 195 2 x 2D scatter plot 185 2 x 2D scatter plot three way PLS 186 4 x 2D scatter plot 186 4 x 2D scatter plot three way PLS 186 general 185 Loading Weights dialog 185 loading
318. n Quality statistics for the following groups of samples 1 All samples 2 Samples which come from the New Equipment 3 Samples which come from the Old Equipment Note If you have category variables or design variables in your table each level of each of these variables will automatically define a group No need to include those leveled variables in the analysis For a designed data table more groups are automatically defined Design samples Center samples Reference samples whenever relevant If there are more than two leveled variables in the data table you have the additional possibility of defining subgroups based on combinations of the levels of at least two of those variables In the Subgroups field click Define to choose the leveled variables you want to use For instance if you choose variables Equipment 2 levels New Old and Week 3 levels 36 37 38 you will get statistics for each of the 6 subgroups New equipment Week 36 Old Equipment Week 36 etc in addition to all pre existing groups All samples New equipment Old equipment Week 36 etc Statistics Dialog Variables Sheet On the Variables sheet you decide on which Variable Set to perform the statistical calculations in the Variable Set field Only continuous variables may be chosen If there are any category variables in your selection they will be automatically excluded from the analysis To de
319. n a Low to High range Continuous variable or takes only a predefined number of values Category variable by selecting the appropriate radio button in the section Data Type Finally you specify the value of the design variable at the different levels of the design in the field Level If a continuous variable is chosen the values are entered in boxes in a graphical view of the design For full factorial designs you are allowed to define more than two levels by clicking Add level If you have chosen data type category the bottom of the dialog shows a list which by default contains two levels Choose a level in the list and fill in its name in the box below the list You may add more levels to the design variable by clicking Add or change the position of a level in the list by using Move up or Move down Note Only two levels are allowed in fractional factorial and Plackett Burman designs Up to 20 levels are allowed in full factorial designs Add Design Variable Dialog Box Behnken Design The dialog has about the same shape as for a factorial design see Add Design Variable Dialog Factorial or Plackett Burman Design but in addition the Center level is displayed The Unscrambler Program Operation The Design Wizard e 105 Dialog Add Design Variable for a Box Behnken design Add Design Variable ID A Name y Data Type Continuous c Levels Add Design Variable Dialog Central Composite Design The dialog
320. n a star are replaced by the sample variable number in plots when the option Use numbers for blank names is selected Uncheck it to display the data point without the number Viewer Colors If you click the button Define Colors on the Viewer sheet you enter the dialog Viewer Colors where you can change the default colors of all elements in the plots to suit your needs for documentation In the field Device you can set up different colors for use on screen and for output on printer Default colors can be restored by selecting 1 of the 2 predefined schemes from the drop down menu in the Scheme field There is one with white and one with black background The Unscrambler Program Operation The File Menu e 93 Q Camo Software AS The Unscrambler User Manual In the Item field you can select the plot item you want to change the color for and set the color using the box to the right of this menu box Color Scheme for Black and White Printing In the Viewer The Unscrambler uses colors that make it easy to distinguish different curves or other objects on screen However when printing out plots on a black amp white printer these colors sometimes appear very similar In such a case a new default color scheme can be used From the Viewer Colors dialog shown below select the following options Device printer Scheme Default Greytones The Viewer Colors dialog offers a Greytones defaults setting Viewer Colors E isto jito Zo
321. n is summed up The height of the bar in the histograms shows how many data points fall within the data range of the bin Hotelling T Ellipse This 95 confidence ellipse can be included in Score plots and reveals potential outliers lying outside the ellipse The Hotelling statistic is presented in the Method References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices Intercept Also called Offset The point where a regression line crosses the ordinate Y axis Interior Point Point which is not located on the surface but inside of the experimental region For example an axial point is a particular kind of interior point Interior points are used in classical mixture designs The Unscrambler Program Operation Glossary of Terms e 305 Leveled Variable A leveled variable is a variable which consists of discrete values instead of a range of continuous values Examples are design variables and category variables Leveled variables can be used to separate a data table into different groups This feature is used by the Statistics task and in sample plots from PCA PCR PLS MLR Prediction and Classification results Levels Possible values of a variable A category variable has several levels which are all possible categories A design variable has at least a low and a high level which are the lower and higher bounds of its range of variation Sometimes intermediate
322. ncel Help 32 Define Sample Set Keep Out of Calculation Select Frozen Calibration Samples Ir eh Validation Method C Leverage Correction Cross Validation Setup IV Uncertainty test Opt PCs C Test Set Model Size Fy Num PCs B IV Center Data Add Start Noise IV Issue Warnings Waming Limits First select the appropriate regression method for your problem and type of data in the Method field The available methods are described below Regression methods Regression Method Description PLS1 Partial Least This method is valid in all situations especially if there is some correlation Squares Regression among your X variables It builds a model with only 1 Y variable PLS2 Partial Least This method is valid in all situations especially if there is some correlation Squares Regression among your X variables It models several Y variables together which is useful if there is some correlation among your Y variables PCR Principal This method is an alternative to PLS1 It runs an MLR on the principal Component Regression components of X Several Y variables can be modeled individually you get results for all Y variables in one run but the results are based on separate models The Unscrambler Program Operation The Task Menu e 269 MLR Multiple Linear This method should only be used when the X variables are not correlated or Regression have only a
323. nd Delete The Unscrambler Program Operation The Modify Menu e 241 e Convert to Category variable e Correct Mixture Components e Convert to Mixture Variables From the Modify menu e Compute e Transform e Sort Samples e Sort Samples by Sets e Sort Variables by Sets e Shift Variables e Reverse Sample order e Reverse Variable order e Swap 3 D Layout e Swap Samples amp Variables e Toggle 3 D Layouts Modify Properties eg This command launches the Properties dialog where the properties of a sample or variable can be viewed and in some cases modified The command is also available by right clicking on a sample or variable number Select the sample or variable for which you wish to view the properties This will launch one of the following dialogs Sample X Properties In this dialog you can change the name of the active sample in the Name field You can also see what kind of sample it is in the Type field All samples in non designed data tables are defined as normal Samples in designed data tables can be of the types listed below Sample types in designed data tables Normal A sample with a normal combination of variable levels Cube A sample with a combination of high and low levels in a design Center A sample with mean values for all variables Star A sample used in Central Composite Designs with extreme values of one design variable and average values of the
324. nd In the 2D 3D Scatter Plot dialog specify a Variable Set if you have selected two samples to plot or a Sample Set if you have selected two variables to plot from the drop down list in the Scope field If necessary press the Define button to define new sets in the Set Editor Use Edit Options to change the layout of the plot Plot 3D Scatter The Unscrambler Program Operation The Plot Menu e 175 The 3D Scatter plot shows three data vectors plotted against each other You enter the 2D 3D Scatter Plot dialog where you can specify the scope of the plot Note When you are plotting from the Editor you must mark the two samples or variables you want to plot before using the Plot command In the 2D 3D Scatter Plot dialog specify a Variable Set if you have selected three samples to plot or a Sample Set if you have selected three variables to plot from the drop down list in the Scope field If necessary press the Define button to define new sets in the Set Editor Use Edit Options to change the layout of the plot Plot Normal Probability The Normal Probability plot shows the deviation from an assumed normal distribution of the data vector It is not possible to plot more than one row or column at a time in this plot Select the sample or variable you want to plot and use Plot Normal Probability You enter the Normal Probability Plot dialog where you choose the scope of your plot Specify a Variable Set
325. nd New Variable Set respectively Note If you are editing a 3 D data table you may also enter one of the following dialogs New Primary Sample Set New Secondary Sample Set New Primary Variable Set New Secondary Variable Set Modify Primary Sample Set Modify Secondary Sample Set Modify Primary Variable Set Modify Secondary Variable Set These dialogs are all identical to New Sample Set and New Variable Set respectively For more details on the fields available in each dialog lookup the one you are interested in hereafter e New Sample Set Dialog p 245 e New Variable Set Dialog p 246 Interaction and Square Effects Selection Dialog When you press the Select button to the right of the Interaction and square effects field in the New Variable Set dialog you enter the Interaction and Square Effects Selection dialog where you can specify combinations of variables cross products of two variables also called Interactions or squares of individual variables to be taken into account in your model The Unscrambler Program Operation The Modify Menu e 249 Dialog Interaction and Square Effects Selection Interaction and Square Effects Selection sl Selected Effects Available Effects OK Name 5 F 7 Cancel Remove gt gt e Add All Add Int Add Squ Remove All The dialog contains two lists Selected Effects to the left and Available Effects to the right The latter lists all available effects with their full names
326. ndidate point 303 cascade 294 category variable 303 category variables 31 in mixture design 107 108 cell locked 31 selection 31 center and reference samples use in significance testing 203 204 center and scale 232 Center and Scale dialog 232 233 The Unscrambler Program Operation Index e 315 center data 260 261 271 272 276 center sample 303 center samples 116 117 use in significance testing 203 204 centering transformation 232 central composite design 98 99 extend 102 centroid design extend 103 Change License dialog 301 change viewpoint 156 Change Viewpoint dialog 156 Choose significance level 149 classification 208 unsupervised clustering 255 Classification dialog 208 278 classify 278 CLASS PA files 63 clear 119 clipboard format 92 93 close all 295 close all windows 295 close current window 294 close file 48 closure 303 clustering 255 Number of Iterations 257 Sum of Distances 257 Clustering Number of Iterations 257 Sum of Distances 257 Clustering dialog 255 colors 35 device 92 93 item 92 93 preset schemes 35 viewer 92 93 column auto adjust 92 component 309 compute 217 Compute dialog 217 compute general 217 condition number 114 Condition Number 115 Configure dialog 201 configure The Unscrambler 90 confoundings 116 resolution III 116 resolution IV 116 resolution V 116 constraint 303 Constraint 303 context saving 91 context sensitive menus 29 continuous variables 31 304 Contour
327. ndows Installer P 09 04 2003 23 33 Bi cetup exe 220KB Application 27 10 2005 18 38 Setup ini 2KB Configuration Settings 27 10 2005 18 38 iS The Unscrambler License Serv 891 KB Windows Installer P 27 10 2005 18 38 Installing The Unscrambler License Server If you have received The Unscrambler License Server installer on CD ROM simply place the CD in your CD player If you have downloaded The Unscrambler License Server from Camo s website double click on file setup exe to start the installation 4 AnInstallShield Wizard dialog opens InstallShield Wizard Preparing to Install The Unscrambler License Server Setup is preparing the InstallShield Wizard which will guide you through the program setup process Please wait Configuring Windows Installer BRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR 5 The InstallShield Wizard for The Unscrambler License Server is launched The Unscrambler Program Operation Network Installation Multiple User Licenses e 9 Camo Software AS The Unscrambler User Manual InstallShield Wizard Welcome to the InstallShield Wizard for The Unscrambler License Server The InstallShield Wizard will install The Unscrambler License Server on your computer To continue click Next InstallShield Wizard License Agreement Please read the following license agreement carefully Press the PAGE DOWN key to see the rest of the agreement MPORTANT READ CAREFULLY The CAMO lic
328. nificance of your X variables in a regression model Variables with uncertainty limits that do not cross the 0 axis are significant 312 e Glossary of Terms The Unscrambler Program Operation Uncertainty Test Martens Uncertainty Test is a significance testing method implemented in The Unscrambler which assesses the stability of PCA or Regression results Many plots and results are associated to the test allowing the estimation of the model stability the identification of perturbing samples or variables and the selection of significant X variables The test is performed with Cross Validation and is based on the Jack knifing principle Unfold Operation consisting in mapping a three way data structure onto a flat two way layout An unfolded three way array has one of its original modes nested into another one In horizontal unfolding all planes are displayed side by side resulting in an OV layout with Primary and Secondary variables In vertical unfolding all planes are displayed on top of each other resulting in an ON layout with Primary and Secondary samples Unimodality In MCR the Unimodality constraint allows the presence of only one maximum per profile Upper Quartile The upper quartile of an observed distribution is the variable value that splits the observations into 75 lower values and 25 higher values It can also be called 75 percentile User Defined Analysis UDA DLL routine programmed in C
329. ning the source data table will also close the Sample Statistics table If you want to manipulate these statistics further you have to select rows or columns of the Sample Statistics table and drag n drop them onto an empty spot of the workplace This will generate a new table which you can freely manipulate edit and save View Sample Statistics 3 D This option is disabled for all 3 D data tables no matter which layout they have Useful tips 1 To display Sample Statistics for an O V table select the whole table then drag and drop it unto an empty spot of the workspace alternatively use File Duplicate As 2 D Data Table You may then display Sample Statistics on the resulting 2 D table The Unscrambler Program Operation The View Menu e 151 The Unscrambler User Manual 2 To display Sample Statistics for an OV table two solutions are possible Duplicating the table as 2 D and displaying Sample Statistics will give you a summary where Primary and Secondary variables are taken into account together If you are interested in Sample Statistics across individual Primary resp Secondary variables here is what you can do Select the sample you want to focus on Choose File Convert Vector to Data Table The resulting 2 D table has as many rows as you had Secondary variables and as many columns as you had Primary variables Display Sample Statistics on that table if you want a summary of each Secondary variable across al
330. ns 191 estimated spectra 191 from the editor 32 from the viewer 33 histogram 149 176 ID 158 important variables 195 leverage 190 line 175 loading weights 195 loadings 185 matrix 177 matrix 3 D 177 Matrix 3 D 177 MCR overview 190 normal probability 176 N plot 176 options 134 PCA overview 177 predicted vs measured 194 prediction 209 regression and prediction 199 regression coefficients 196 regression overview 191 regression overview three way PLS 192 residuals 186 response surface 199 response surface for mixture process 200 201 response surface for mixtures 200 201 response surface for process 200 201 response surface overview 202 RMSE 180 RMSEP 178 sample outliers 181 scores 183 scores and loading weights 182 scores and loadings 181 stability 160 statistics 158 205 uncertainty limits 161 variances 178 variances and RMSEP 178 X and Y variance 179 X or Y variance 178 X or Y variance three way PLS 179 X1 and X2 variance 179 information 34 normal probability 307 overview 33 34 point details 34 predefined result 33 34 remember settings 92 93 PLS 268 269 PLS 1 268 269 308 PLS2 268 269 308 point details on plot 34 point names 154 predefined result plots 33 34 predict 282 predicted vs measured 194 Predicted vs Measured dialog 194 predicted vs reference 209 210 predicted with deviation 209 210 prediction 209 308 predicted sheet 209 residuals leverage sheet 210 Prediction Three Way PLS dia
331. ns are useful when you need to examine all interactions and can afford to do all the experiments You enter the following dialogs Define Design Variables Define Non design Variables Design Details Randomization Details Last Checks Extend to Central Composite Design Use the central composite design when you go from the screening phase to the optimization phase You enter the following dialogs Define Design Variables Define Non design Variables Design Details Randomization Details Last Checks Note Category variables must be removed before the design is extended to a central composite design The Unscrambler Program Operation The Design Wizard e 101 Design Wizard Extend a Central Composite Design The only extension option for Central Composite Designs is to add center and or reference samples and replicates Center samples and replicates are added to increase the reliability of the model Add reference samples to see how your design samples compare to them You enter the following dialogs Design Details Randomization Details Last Checks Design Wizard Extend a D optimal Non mixture Design You can extend a D Optimal Non mixture Design by adding reference samples center points or replicates Tick the radio button Add Center and Reference Samples or Replicates and press Next You will enter the following dialogs Design Details and Last Checks Design Wizard Extend a D optimal Mixture Design You can extend a D Opt
332. nscrambler Activation dialog opens The Unscrambler Activation Unscrambler Gees sg machine code 21D 36605 Jptai Unscrambler IR activation key ctivate Currently The Unscrambler v9 5 is not activated c Click on the Obtain button to request your Unscrambler activation ee key on line If you have no internet connection contact your Camo office and provide us with your Unscrambler machine code If you have Internet access Click the Obtain button to request your Unscrambler activation key online You will be brought to the activation key request form on Camo s website When you submit your request your Unscrambler activation key will be automatically sent to your e mail ID Note Your e mail ID is the means of identifying you as a registered user Make sure to type in the same e mail address in the form as the one you registered as an Unscrambler licensee to receive a permanent activation key If you are not a registered Unscrambler licensee a 30 day activation key will be sent to your e mail ID 6 e Installation and Configuration The Unscrambler Program Operation If you do not have Internet access Contact Camo and give your Camo consultant the Unscrambler machine code that is displayed on screen Your consultant will provide you with an activation key in accordance to your license The Unscrambler Actsting 21D866C6 Unscrambler activation key Currently The Unscrambler v9 5 is not activate
333. nt file formats which later on can be opened in the application of your choice A Special Case Transferring Table Plots Table plots e g ANOVA table are nothing but characters and numbers displayed in a graphical interface Thus if you copy a table plot and paste it into another application you will get the following results e Ina spreadsheet e g Excel A table with row and column names e Ina word processor e g Word A tabulated series of words and values which can easily be converted into a table see section Transferring Your Data to a Word Processor Import Export in Practice Menu options for import and export with various formats 42 e Communication with Other Software and Instruments The Unscrambler Program Operation How To Import Results or Data 2 D With the File Import options you can import a data table from another program as well as results from an Unscrambler model Once you have made all the necessary specifications in the Import and Import from Data Set dialogs a new Editor which contains the imported data will be created in The Unscrambler Read more about the File Import menu options in chapter File Import p 50 How To Import Results or Data 3 D With the File Import 3 D options you can import a three way data table from another program as well as three way result arrays from an Unscrambler model Once you have made all the necessary specifications in the dialogs a new Editor
334. nu e 207 Dialog Statistics Cross Correlation sheet Statistics Plat Plot Type Iw Table Cross Cortelation Iw Matrix The default layout for the matrix plot of cross correlations is Bars as shown in the plot preview Once the plot is displayed after you clicked OK you may change the layout if you wish by using Edit Options or or Ctrl L Plot Classification Plots that are specific to classifications are plotted from the Classification dialog The different plots available in the dialog are displayed below Classification plots Type Description Shows the orthogonal distances from the new objects to two different models at the Coomans same time Si vs Hi Shows the limits used for classification both distance to model residual standard deviation and leverage distance to model center Si HO vs Hi The same as Si vs Hi but the distance measures are relative to the average distance in the model Classification Table Shows compressed numeric results for each classified sample Model Distance Shows the distances between different models Shows how much the different variables contribute to separating different models Modeling Power Shows how much the variables contribute to the model All of these plots except the Classification Table require that you select the model for which you want to see the Classification results from the drop down menu Data From Mod
335. number of times to find improved weights The Unscrambler Program Operation The Modify Menu e 227 The Unscrambler User Manual Modify Transform Noise You can introduce additive and or proportional noise into the selected data range This may be useful to see how sensitive your model is to noise in the data This transformation has no specific row or column orientation that is to say you can compute the new value of a cell independently from its neighbors In the Noise dialog select the sample and variable sets that define the matrix you want to correct in the Scope field If necessary press Define to enter the Set Editor and define new sets Then in the Parameters field specify the level of proportional noise e g 5 and the standard deviation of the additive noise you want to add Modify Transform Derivatives Differentiation i e computing derivatives of various orders is a classical technique widely used for spectroscopic applications Some of the information hidden in a spectrum may be more easily revealed when working on a first or second derivative It is a row oriented transformation that is to say the contents of a cell are likely to be influenced by its horizontal neighbors Three algorithms are available for computing derivatives 1 Norris Gap 2 Gap Segment 3 Savitzky Golay Modify Transform Derivatives Norris Gap This method computes derivatives of up to the 4 order It has the advant
336. o the contents of a data table Locking data sets is a recommendation of the 21 CFR Part 11 guidelines from the FDA The Unscrambler Program Operation The File Menu e 83 Once a file is locked the usual Editor operations performed with the Edit and Modify menus copy paste replace transform etc are disabled as well as direct data input and import of new data In the General sheet of the Properties dialog the file is shown as Read Only and the Summary field at the bottom mentions This file is currently locked Note Locking a file does not prevent you from running analyses on the data Those analyses will automatically be kept track of in the Log sheet of the Properties dialog as usual To enable changes in the data table use Feil Fant ikke referansekilden as described hereafter File Unlock This command enables changes to the contents of a data table that was previously locked see Feil Fant ikke referansekilden above In the General sheet of the Properties dialog the file is shown as Read Write and the Summary field at the bottom mentions This file is currently unlocked File Properties Information about the contents of the data or result file in the current window is shown in the Properties dialog The shortcut button see above on the Toolbar takes you directly to the Notes sheet of the dialog The Properties dialog contains up to 5 out of a total of 7 available sheets which
337. o the specifications in the Warning Limits dialog which you enter by clicking the Warning Limits button Note Remember that warnings are time and memory consuming Turn them off if you do not intend to use them Invalid Sample or Variable Range In all analysis dialogs Statistics PCA Regression etc you have the possibility to specify samples or variables to be kept out of the calculations The range for those samples or variables may be entered either using an Editor through the Select button or directly Upon direct entry you may get an error message if the range you specified is invalid with respect to the samples or variables present in the Sample set or Variable set to be analyzed Error message upon invalid sample or variable range The Input Yalue Is not Accepted Change Value to OK Cancel Should this dialog show up type in a new range within the range of samples or variables present in the current Sample set or Variable set If you do not remember the correct range click Cancel to return to the analysis dialog and use the Select button to select a valid range Set Weights Dialog PCA PLS and PCR analyses depend on the relative standard deviation of the variables For PLS2 both X and Y variables can be weighted so that only the relative differences among the variance of the X variables and the relative differences among the Y variables influence the model When you press the Weights but
338. of plot is activated The options are listed below General plot options available for histograms Option Description Gria Displays a grid allowing you to see the data points coordinates more clearly Piot Header Draws the title of the plot and axis names Options Dialog General Sheet for Line and Matrix Plots Accessed from Edit Options for a Line plot or Matrix plot this sheet lets you define the plot layout The Unscrambler Program Operation The Edit Menu e 139 Dialog Options General sheet for Line plots and Matrix plots Options h x General r Plot Layout a E Equal Scale C Bars r SS e Cancel urve C Accumulated Bars E Vertical line E C Symbols landscape E Connection Line Contour I Plot Header C Wap Labels Layout Name uada Number Venable Level Sie fication Name Orientation Position on File GG Horizontal None C Vertical Options on the General sheet apply to all the data points in the plot The check boxes in the upper right field are used to closer characterize the plots and are enabled and disabled according to which type of plot is activated The available boxes and their characteristics are listed below General plot options available for Line plots or Matrix plots Option Description Grid Displays a grid allowing you to see the data points coordinates more clearly Curve ID Identifies each curve in the plot w
339. oftware from for the CAMO software product The software products covered by this agreement is The Unserambler R T he Unscrambler A Designer The Unscrambler R MYA and The Unscrambler License Server hereafter referred to as the Program and includes computer software and associated media and printed materials and may include on line or electronic documentation By installing copying or otherwise using the Program you agree to be bounds by the terms in this license agreement If you do not agree to the terms in this license agreement promptly return the unused software to the place from which you Do you accept all the terms of the preceding License Agreement If you choose No the setup will close To install The Unscrambler you must accept this agreement InstallShield lt Back 7 Select a destination folder for the The Unscrambler Program files The Unscrambler Program Operation Standalone Installation 1 User License Desktop Installation e 3 InstallShield Wizard Choose Destination Location Select folder where Setup will install files Setup will install The Unscrambler in the following folder To install to this folder click Next To install to a different folder click Browse and select another folder Destination Folder C Program Files The Unscrambler Leet alle sld InstallShield 8 Select a destination folder for the The Unscrambler DATA files InstallShield Wizard Choose
340. og Instrument Sheet This sheet exists only for data files When data are imported from e g NSAS and JCAMP there are some parameters which do not fit into the Unscrambler data format These parameters are stored in the Instrument sheet and are used for file export at a later stage This is only possible for the data tables and not for results so if you need instrument parameters when you export a model set these parameter in the data file then perform the calculation and the export of the model The Unscrambler Program Operation The File Menu e 87 Camo Software AS The Unscrambler User Manual Dialog Properties Instrument Sheet SPC data Properties x General Notes Log Instrument Constraints Instrument parameters describe settings used by an instrument and or measurement process resulting in this data table Parameter Value EO Original_Sample_Name n ftirsdatasd0958 m002 Remove All Comment NMF micro ATR Si MCT Atplus 309F 2 Requested_Scans_per_Scanset 1000 Number_of_Scansets 10 Scan_type Absorbance Background_File n ftirSback met micro_atr_si_C Resolution 2 Igram_Symmetry Single Sided Sampling UDR 2 Gain_Range_Radius 40 Gain_Range_Slope 14 675407 EI Cancel Apply You may enter your own instrument parameters if additional information is needed The instrument parameters are explained in more detail in the Technical References chapter which i
341. og Response Surface Response Surface x Samples variables Y variables Variable Set Define EC Keep Out of Calculation Select M Issue Wamings Warning Limits Warnings In the bottom field of the dialog you define whether warnings should be issued by ticking or un ticking the Issue Warnings box Ticking it will enable the Warning Limits button which launches the Warning Limits dialog Response Surface Dialog Samples Sheet Choose which samples to use in your calculation in the Sample Set field and enter any samples which you do not want to include in the analyses in the Keep Out of Calculation field For more details you may lookup the description of a similar dialog in Statistics Dialog Samples Sheet Response Surface Dialog X variables Sheet and Y variables Sheet In the Variable Set field you choose which Variable Set to use as X and Y respectively in the calculations You can either choose an existing Variable Set from the drop down list or define a new in the Set Editor which you enter by clicking Define The computations can only be done for continuous variables If there are category variables in the data selection you must choose their level before computation starts The category variable list box lists all the category variables You will have to select one level for each variable by highlighting it in the list and then selecting one of its levels The Unsc
342. oints 106 star points distance to center 311 star samples 312 distance to center 311 start noise 260 262 271 273 startup dialog 23 26 Startup dialog 26 statistics 205 253 compressed sheet 206 cross correlation sheet 207 general sheet 206 group sheet 206 scatter sheet 207 Statistics dialog 253 status bar 30 173 sub views 292 293 Sum of Distances 257 supervisor responsibilities 22 swap 3 D layout 239 swap primary samples and secondary samples 240 swap primary samples and variables 241 swap primary variables and samples 240 swap primary variables and secondary variables 240 swap samples amp variables 239 pri sample amp sec sample 240 pri sample amp variable 241 pri variable amp sample 240 pri variable amp sec variable 240 sec variable amp sample 240 241 swap secondary variables and samples 240 241 system configuration supervisor 23 system setup 90 System Setup dialog 90 System Setup dialog supervisor 23 T target line 159 task analysis of effects 278 classify 278 clustering 255 extract data from marked 288 extract data from unmarked 288 MCR 267 multivariate curve resolution 267 PCA 258 predict 282 284 principal component analysis 258 recalculate with marked 287 recalculate with passified marked 288 recalculate with passifiedun marked 288 recalculate without marked 288 regression 268 273 response surface 277 statistics 253 User defined Analysis 286 test samples 146 312 test set 260 test se
343. om past experience There is no such thing as correct limits limits are only filters used to highlight the most extreme results of each test The actual value of a useful limit may vary from one application to another If you get too many warnings raise the limit On the other hand if you get few warnings even though you know there are outliers decrease the limit Limits of 4 5 are usual for spectroscopy data Note All limits do not apply to every analysis Only those that apply to the analysis you are about to perform are shown in this dialog The options in the Warning Limits dialog are listed and described below Warning limits Limit Description Leverage Limit Objects with high leverage have a large influence on the Ratio of Calibrated to Validated Explained Variance Ratio of Calibrated to Validated Multiple Correlation Ratio of Validated to Calibrated Explained Variance Indicates an unstable structure in the model or that outliers may be present model and may be dangerous if erroneous Calibration set may lack representativit Calibration set may lack representativity Ratio of Validated to Calibrated Multiple Correlation Sample Outlier Limit Calibration Validation set may lack representativity Some samples are badly fitted Sample Outlier Limit Validation Statistical Condition Number Total Explained Variance Some samples are badly fitted The X matrix is i
344. on Setup 265 Text Properties 144 145 Text Warning List 281 Tip of the Day 300 Transform Baseline 231 Uncertainty Test 265 User Defined Analysis 286 User Defined Transformation 234 Variable Properties 104 244 Variable Statistics 153 Variance 282 Variances and RMSEP 178 Viewer Toolbars 173 Warning Limits 146 266 X Y Relation Outliers 193 dialogs 36 distribution 305 dockable views 35 D Optimal algorithm 114 D Optimal Design 89 drag n drop 41 50 67 Drag n Drop 67 draw item 143 line 144 text 144 duplicate 80 as 2 D Data Table 81 as 3 D Data Table 81 as modified design 81 as non design 81 318 e Index The Unscrambler Program Operation Incerramhbilaear ar Ma T SCrampic JSel lanual E edge center point 305 edit add plot 134 adjust width 133 append 125 append category variable 126 append mixture variables 126 clear 119 convert to category variable 129 convert to mixture variable 131 Convert to Mixture Variable 131 copy 119 copy all 149 correct mixture components 129 cut 119 delete 126 delete item 144 fill 120 fill missing 120 find replace 127 go to 131 insert 122 insert category variable 124 insert draw item 143 insert mixture variables 125 item properties 144 mark 146 options 134 paste 119 select all 133 select bars 149 select samples 132 select variables 133 split category variable 131 edit set 246 3 D data 246 editor 29 30 33 effects 203 effect details 205 overview 203 respon
345. only centered models are available for selection You select which models to use in the classification by pressing Add Model This takes you to the Get Model dialog where you can select the appropriate model The number of X variables must be the same in your new data and all models you use to classify it SIMCA classification is only possible based on PCA PCR and PLS models In the case of PCR and PLS models only the X part of the model will be taken into account Select each model in turn and specify how many PCs to use with the spin button The Number of PCs to Use should be decided individually for each model you use in the classification Your choice updates the number of PCs for this model in the classification dialog A warning message is issued if you try to use more PCs than the optimal number suggested by The Unscrambler Press Variance to inspect the calibration and validation variances of the selected model This takes you to the Variance Dialog where the variances are shown for each PC in the selected model Toggle between residual and explained variance by pressing the button Use the Pretreat button to pretreat the model this takes you to the Select transformations Dialog where you may choose one or several of the pretreatments that were applied to the calibration data prior to building the class models For an explanation of the abbreviations referring to various transformations see the Technical References chapter which is
346. ons The Unscrambler Program Operation Dialog Line Properties Line Properties Style EME wan fr 2 Cancel Color Cam vr DEDA M O00 ae S ll CL LISS HANT Make a choice from the Style drop down list Tune Width up or down Select a suitable color black is default from the Color field Text Properties Dialog This dialog allows you to choose the properties of a text inserted in your plot as draw item Dialog Text Properties Text Properties Font Arial X Cancel Siz Small Nom Large Help Align Left Center C Right e Top Center Bottom Color WIR ll DAND M OOT eS eee H BEBO E Type in or correct your text in the Text field Make a choice from the Font drop down list Make a choice of Size and horizontal vertical alignment using the corresponding radio buttons Select a suitable color black is default from the Color field The Unscrambler Program Operation The Edit Menu e 145 Edit Mark Mark objects samples or variables you want to focus on in your interpretation The menu option Edit Mark has 8 sub options Ze e Outliers Only H o e Test Samples Only s e Evenly Distributed Samples Only e Significant X Variables Only Al e With Rectangle 5 e One By One E e Reverse Marking il e Unmark All Edit Mark Basic Notions A typical use of this command is to mark extreme samples in a score plot and investigate the behavior of those sa
347. ons let you switch between absorbance and reflectance data and transform reflectance data into Kubelka Munk units Dialog Spectroscopic Transformation Spectroscopic Transformation x m Scope OK Sangli EC All Samples 25 e Define Cancel Variables Help All Variables 262 e Define Type Reflectance to absorbance or Transmittance to absorbance Reflectance to Kubelka Munk Unit Conversion C Wavenumber em 1 to Nanometer nm Nanometer nm to Wavenumber om 1 FE Inelude em or mmm unittvarable name In the Spectroscopic Transformation dialog select the sample and variable sets that define the matrix you want to transform in the Scope field If necessary press Define to enter the Set Editor and define new sets Then select the transformation type in the next fields Three types of transformations can be performed on the data values using the options from frame Type 224 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS 1 Absorbance to reflectance Absorbance to transmittance 2 Reflectance to absorbance Transmittance to absorbance 3 Reflectance to Kubelka Munk As for the frame Unit Conversion it allows you to convert the units for the variable names by shifting from Wavenumber cm 1 to Nanometer nm and vice versa Tick the box at the bottom if you wish to include cm 1 or nm in unit variab
348. ortance of the sensory property Thickness becomes much more obvious when the correlation loadings ellipse is shown Usual Loading plot REDNESS COLOUR 0 3 0 03 0 6 0 9 RESULT4 X expl 56 21 Y expl 82 13 Correlation Loading plot displaying the two ellipses The Unscrambler Program Operation The View Menu e 161 PC2 Correlation Loadings O9 1 0 0 5 o KE 1 0 RESULTS X expl 56 21 Y expl 82 13 Correlation loadings are computed for each variable for the displayed Principal Components In the figure above the outer ellipse is the unit circle and indicates 100 explained variance The inner ellipse indicates 50 of explained variance For information on the calculation of correlation loadings we refer to Westad 1999 see chapter Bibliographical References Note The Correlation Loadings plot can only be displayed if the currently active plot is a two vector Loading plot It does not apply to a plot of X Loading Weights and Y Loadings nor to a line plot or a three vector plot of Loadings View Source The toggle switches in this option give you an easy way to change the source data for the current plot without going through the Plot menu For instance you can display Calibration variances only Validation variances only or both There are 13 options How many of them are enabled depends on the kind of plot you are in The options enabled in each type of plot are listed below Available sources
349. ot Response Surface Response Surface Dialog In the Plot field in the Response Surface dialog you can choose whether or not to include a table which shows the measured and predicted values of the Y variable plus the standard error of the prediction for each sample This will automatically place the plot itself in sub view 2 and the table in sub view 3 If you have not included the table plot you can change the size of the plot by double clicking in the preview screen in the dialog Dialog Response Surface Response Surface ES OK Cancel Help ud WE Plot 1 l Response Surface P Include Tables General p Ze rte tivture Process Process Axis Range 1 La x 0 0 to 57 5 2 Java x 18 3 to 75 8 3 Flour B D 15 8 to 73 3 Configure In addition the Response Surface dialog contains two sheets 1 General 2 X variables Response Surface Dialog General Sheet In the General sheet three fields must be configured Layout Y variable and Components In the Layout field you make your choice between the two plot types Then define the Y variable you want to investigate in the Y variable field Finally set the number of Components in this field Response Surface Dialog X variables Sheet The X variables sheet lets you select the type of plot you need and choose relevant X variables Three types of plot are available corresponding to the following situations 1 Mixture
350. ots will appear with the same layout as they had in the Viewer Edit Select Bars Select how many bars are to be used in histogram plots The Unscrambler calculates the value limits for each bar The number of bars must lie within the limits given in brackets Adding more bars gives you more details about the distribution of the data points Choose Significance Level 5 e This command is accessed from the toolbar only when viewing classification results The tool is active for the following plots e Classification table e Cooman s plot e Si vs Hi e Si SO vs Hi Choose the desired significance level either None or between 0 1 and 25 for the classification results by clicking on the arrow and selecting desired level from the list If you tune up the significance level e g from 5 to 10 this will narrow down the class limits more samples are rejected If you tune it down e g from 5 to 1 this will widen up the class limits fewer samples are rejected If you choose None on the Cooman s plot Si vs Hi or Si SO vs Hi this will remove any significance limits form the plot Significance Limits None In the case of 2 D scatter plots Cooman s plot Si vs Hi Si SO vs Hi an additional level is available Select None from the drop down list if you wish to display the plot without any significance limits The Unscrambler Program Operation The Edit Menu e 149 The Unscrambler User Manu
351. otting the different results you have made including results that are not available as predefined model plots In the General View select Plot on the Menu Bar and select which kind of plot you want to make This will launch a plot dialog where you must specify which data or result file you want to plot samples variables from Details e Plot Line Plot one vector at a time as line see p 211 e Plot 2D Scatter Plot 2 vectors against each other see p 212 Plot 3D Scatter Plot 3 vectors against each other see p 213 Plot Histogram Plot one vector as histogram see p 214 Plot Matrix Plot one or several vectors together as matrix landscape contour etc see p 215 Plot Normal Probability Plot one vector as N plot see p 215 The Window Menu Window Copy To This option copies the plot in the current sub view Source into another sub view destination Seven different sub views are available each of them may have some contents even though only a maximum of four are displayed at the same time The sub views are organized into three sub view systems 1 One sub view the Viewer window contains only one sub view number 1 which uses the whole window space 2 Two sub views the Viewer window is divided into two horizontal sub views numbers 2 and 3 each of which uses exactly one half of the window space 3 Four sub views the Viewer window is divided into four sub views numbers 4 to 7 each of whi
352. ough the successive stages of a longer operation by means of Next and Back buttons Dialogs When you are working in The Unscrambler you will often have to enter information or make choices in order to be able to complete your project such as specifying the names of files you want to work with or the sets which you want to analyze or how many PCs you want to compute This is done in dialogs which will normally look something like the one pictured below Unscrambler dialog et erf HEES Drop down list Select a tab to see more options amples Variables Sample Set All Sample Keep Out of Calculation 13 20 Select Frozen Calibration Samples Validation Method List of values of ranges Select C Leverage Correction Cross Validation Uncertainty test al Test Set Radio button Opens a new dialog Model Size Full e Num PCs 3 IV Center Data Add Start Noise IV Issue Warnings Warning Limits This particular dialog is the one you enter when you want to run a Principal Components Analysis on your data Items that are predefined such as sets file filters etc are selected from a drop down list Ranges of samples or variables are entered as shown in the Keep Out of Calculation field in the figure You can use a comma to separate two items in a field and a hyphen to specify the whole range between two values Opt
353. our table Note You cannot append Mixture variables to a 3 D table Edit Delete x or lt Del gt This option only applies when you have selected one or more variables or samples and deletes the selected section s 126 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Variables or samples with locked cells cannot be deleted Any previously defined sets are adjusted for the deleted range You will be warned before deletion occurs if you have specified this option in the System Setup dialog Edit Delete Mixture Variable This menu option can be used to delete mixture variables It will fail for any variable selection that would leave only one or two mixture variables in the table after the deletion Edit Delete 3 D If your table has an OV layout you may delete either Primary or Secondary variables as shown below Samples can be deleted as in 2 D data Dialog Delete Variables ov Delete Yariables x What would you like to delete Primary variable 2 removing 5 columns ES Secondary variable 3 removing 6 columns Help If your table has an O V layout you may delete either Primary or Secondary samples as shown below Variables and Category variables can be deleted as in 2 D data Dialog Delete Samples O V Delete Samples x What would you like to delete Primary sample 2 removing 5 rows
354. own File New 3 D data table layout OV Specify the size of the table by filling in the fields Primary variables Secondary variables and Samples Type in the desired number or use the spin button to tune the value up and down Fill in the names of your samples by double clicking on the left margin where sample names are displayed Fill in the names of your variables by double clicking on the upper margin where variable names are displayed Variable names consist of two parts Primary Var Name Secondary Var name The two parts are separated by a hyphen If you try to enter something else in the Variable Name field an error message will be displayed prompting you for a correct entry see the Invalid Format dialog See the figure below for an example of the ON layout OV2 data table in the Editor DI My OV2 data Bisi x 46 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS File New 3 D data table layout O V Specify the size of the table by filling in the fields Variables Primary samples and Secondary samples Type in the desired number or use the spin button to tune the value up and down Fill in the names of your samples by double clicking on the left margin where sample names are displayed Sample names consist of two parts Primary Sam Name Secondary Sam name The two parts are separated by a hyphen C 71 If you try to enter something else
355. p Samples amp Variables Pri Variable amp Sec Variable ec eee ee eeeeeeee 238 Modify Swap Samples amp Variables Pri Variable amp Sample cece eeeeeeeeeeeeeeees 238 Modify Swap Samples E Variables Sec Variable amp Sample cece eeeseeeeeeeeees 239 Modify Swap Samples amp Variables Pri Sample amp Sec Sample eects 239 Modify Swap Samples amp Variables Pri Sample amp Variable eee ee ereere 239 Modify Swap Samples amp Variables Sec Sample amp Variable lees eeneeees 239 Modify Toggle 3 DLawouts cece ccceeceeesescnecescseeseceecseesecessesseceeessasseceeseasseceseaeaeenes 240 Modify Wt EE 240 Modify WE 241 Modify e En EE 242 Modify Tsay OU ET 243 Modify Edit Set c c sssscsses ssacsesvesssesveceassoveds snnsncetesatsaencosesvantesszensbsesezendptesesensnsetsinsnevevstnas 244 New Sample Set Modify Sample Set and New Variable Set Modify Variable Set 249 Interaction and Square Effects Selection Dialog 249 Select Samples Select Variables Dialog 250 The Task Men EE 253 Taske AS UAL TEE 253 Progress DIALOG eiere annei Eeer Are 254 ASK EE 255 ASK EE 258 Invalid Sample or Variable Range 262 Set Weishts E TEE 262 Cross Validation Setup Dialog cies esceesee ceeeeeeceeesceeeeaesesseeseceeeessessseseeaesaseaeeaaeeees 264 Test Set Validation Setup Dialog 265 Uncertainty Test Dialoger onn oii dE EE 265 Warming Limits Dialog sessirnar ne E Ee ERE EEOAE eegene 266 Task MCR EE 267 EE NE
356. p page 144 Edit Insert Draw Item Text Allows you to insert a text element into your plots Place the cursor at the location where you want the text to appear click with the left mouse button This takes you to the dialog Text Properties where you select the properties of the text Type the text in the Text field and click OK The text can be changed at a later stage by marking the text item Press the lt Alt gt key and click on the text and selecting Edit Item Properties For details on the Text Properties dialog lookup page 145 Edit Delete Item The option is only enabled if you have marked a drawing item line text etc in a plot by pressing the lt Alt gt key and clicking on the item with the left mouse button The item can then be deleted by way of this menu choice Edit Item Properties E If you have marked a drawing item line text etc in a plot by pressing the lt Alt gt key and clicking on the item with the left mouse button the properties of the drawing item such as style width and color can be changed in the Item Properties dialog The contents of the dialog depend on the type of draw item see details below The dialog can also be reached by pressing the lt Alt gt key and double clicking on the draw item with the left mouse button Line Properties Dialog This dialog allows you to choose the properties of a line inserted in your plot as draw item 144 e The Unscrambler Menu Opti
357. ple Residuals A plot of residuals versus variable number for selected samples useful for detecting outliers and models with bad fits The options available beside and below the Plot field may vary depending on the type of plot you have chosen All possible options are described hereafter e In the Y residuals field choose between raw residuals or studentized residuals e Inthe Sample field select for which sample you want to display Sample residuals e Inthe Primary X Variable Secondary X Variable and Y Variable fields choose the appropriate variable s e Inthe Samples field to the right available for Y residuals vs Predicted Y and Variance per Sample select Calibration Validation or both types of Samples e Inthe X Variable Mode field to the right available for Variable Residuals with additional choice X select Primary Secondary or both types of X variables Selecting both will produce a line plot showing all samples for the selected X1 X2 Selecting only one mode will produce a matrix plot combining all samples with all variables from the other mode e Inthe Variables field to the bottom right choose either X variables Y variables or both e Inthe Components field at the bottom choose the number of components for which you want the results displayed The All button allows you to plot results for all PCs simultaneously but the plot may be quite messy Residuals Dialog Outliers Sheet From the Outliers sheet
358. ple mode Secondary samples are nested within each Primary sample Primary Variable In a 3 D data table with layout OV this is the major Variable mode Secondary variables are nested within each Primary variable Principal Component PC Principal Components PCs are composite variables i e linear functions of the original variables estimated to contain in decreasing order the main structured information in the data A PC is the same as a score vector and is also called a latent variable Principal components are estimated in PCA and PCR PLS components are also denoted PCs Process Variable Experimental factor for which the variations are controlled in an experimental design and to which the mixture variable definition does not apply Proportional Noise Noise on a variable is said to be proportional when its size depends on the level of the data value The range of proportional noise is a percentage of the original data values Random Order Randomization is the random mixing of the order in which the experiments are to be performed The purpose is to avoid systematic errors which could interfere with the interpretation of the effects of the design variables Reference Sample Sample included in a designed data table to compare a new product under development to an existing product of a similar type The design file will contain only response values for the reference samples whereas the input part the design part
359. plots 1 and 3 first Scores plot and first Loadings plot and the two axes common to plots 2 and 4 e Inthe Samples field select whether to plot Calibration and or Validation samples e In the Loadings field select whether to see X or Y loadings or both Scores and Loadings Dialog Bi plots Sheet Here you can specify one plot which displays scores and X loadings projected onto the same set of 2 PCs You only need to specify which PCs to plot in the Components field Plot Scores and Loading Weights This is a quadruple plot of scores and loading weights from a three way PLS model along two selected PCs This plot is useful for interpreting the relationship between X variables and Y variables while checking at the same time where the samples project This may tell you about possible outliers or influential samples and give you some hints as to which variables explain sample locations There are three different ways to plot Scores and Loading Weights 1 Two plots 2 Four plots 3 Bi plots These are represented by different sheets in the Scores and Loading Weights dialog Dialog Scores and Loading Weights Scores and Loading Weights Two Plats Four Plots Bi plots ae SH camo Ke Samples Plot I Calibration Scores I Validation Loading Weights Loadings Mode Components Primary 7 C Suggested no of components is 4 Secondary Vector 1 f1 Loadings Vector 2 f2 182 e The Unscr
360. ptions The Unscrambler Program Operation Add Center and Reference Samples or Replicates Center samples and replicates are added to increase the reliability of the model Add reference samples to see how your design samples compare You enter the following dialogs Design Details Randomization Details Last Checks Extend to Central Composite Design Use the central composite design when you go from the screening phase to the optimization phase You enter the following dialogs Define Design Variables Define Non design Variables Design Details Randomization Details Last Checks Note Category variables must be removed before the design is extended to a central composite design Design Wizard Extend a Fractional Factorial Design In this dialog you have the choice between the following 4 extensions Delete Design Variables Design variables that are found to be insignificant can be removed from the data table before you go on with further analyses You enter the following dialogs Define Design Variables Define Non design Variables Design Type Design Details Randomization Details Last Checks Add Center and Reference Samples or Replicates Center samples and replicates are added to increase the reliability of the model Add reference samples to see how your design samples compare You enter the following dialogs Design Details Randomization Details Last Checks Extend to Full Factorial Design Full factorial desig
361. ptions The Unscrambler Program Operation Validation methods validation method Description Leverage Correction Quick and dirty Used for a rough initial estimate Do not use it for final validation as the leverage correction validation results are often too optimistic Cross Validation Simulates test set validation May be time consuming Allows you to use Martens Uncertainty test Test Set Validates the model against separate samples If you choose the method Cross Validation the Setup button next to it will be enabled Press it to enter the Cross Validation Setup dialog where you can further define your cross validation settings Tick in the box for Uncertainty Test only active if Cross Validation has been selected if you wish to take advantage of cross validation segments for assessing the stability of your results Martens Uncertainty Test Tick the button Hi access the Uncertainty Test dialog where you can set up the number of PCs to use in the Uncertainty Test this is not the same as the number of PCs computed in your analysis If you choose the validation method Test Set press Setup to enter the Test Set Validation Setup dialog where you can further define your test set validation settings Regression Dialog Miscellaneous Options PLS and PCR In the bottom of the dialog you can find general options for your analysis 1 Model Size Number of PCs to be computed Center Data Add Star
362. r Windows user guide for more information Startup The Main Window When you start The Unscrambler you enter the Main Window as seen in the figure below The Unscrambler main window K The Unscrambler File View Results Help Titlebar Statusbar For Help press F1 The Menu Bar All operations in The Unscrambler are performed with the help of the menus and options available to you on the Menu bar The figure above shows the default menus which are enabled when you load The Unscrambler which all in turn contain several sub menus and options When you have an Editor or a Viewer open more menus will be available Some sub menus and options may be invalid in a given context these are grayed out Context Sensitive Menus The Unscrambler also features so called context sensitive menus You access these by clicking the right mouse button while the cursor rests on the area on which you want to perform an operation The context sensitive menus are a kind of short cut as they contain only the options which are valid for the selected area which will save you the work of having to click your way through all the menus on the Menu bar The Unscrambler Program Operation Working with Windows e 29 Camo Software AS The Unscrambler User Manual The Toolbar The Toolbar buttons give you shortcuts to the most frequently used commands When you let the mouse cursor rest on a toolbar button an short explanation of its function appears
363. r example 5 rows and 4 columns in the figure below or they will not be displayed in the pull down choice menus for sample and variable names Import Matlab dialog Import Matlab X Select matlab matrices to be imported as data sample names and variable names To disable import of names select the none item Cancel Data Data 5 4 El Help Sample names RowLabels 5x16 DI Variable names ROCHA ddi Matlab variables representing sample and variable names must be character arrays What Cannot be Converted The following cannot be imported from Matlab to The Unscrambler 1 Matrices containing imaginary numbers Cells arrays Structures ee Zen NS Sparse matrices File Import MVACDF In The Unscrambler 7 6 a new import format was implemented to answer the specific needs of one of our major Unscrambler users This format is called MVACDF and is accessible from the menu File Import MVACDF Note If you also collect your data in a particular format and would like The Unscrambler to be compatible to this format please take contact with us 62 e The Unscrambler Menu Options The Unscrambler Program Operation File Import CLASS PA amp SpectrOn This option lets you import data from Guided Wave files with the following formats asc scn and autoscan Dialog CLASS PA amp SpectrOn Import CLASS PA amp SpectrOn Import Look in C My Documents Unse805 DATA Browse OK C
364. r for help regarding how to formulate your query File Import JCAMP DX You may import one or several JCAMP DX files into The Unscrambler s Editor More details regarding JCAMP files are given in the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices Select the files you want to import from the file list in the dialog JCAMP DX Import or use the Browse button to get a list of available files The different files must have the same number of X variables and the same contents in the Y matrix to allow simultaneous import All files that do not conform to the contents of the first selected file are grayed out The Unscrambler Program Operation The File Menu e 59 Dialog JCAMP DX Import JCAMP DX Import Lookin C Spectroscopy Instrument files Browse Dimensions Cancel File name xvas Y vars Samples Help F Include Sample number JCAMP file information Title C Program Files Sland2 002 Origin Yokogawa Electric Corporatio Owner a Data formats Pons sSsSsS S You can select a group of files by pressing the lt Shift gt key while you click on the files you want to import Option Include Sample number will import the sample number together with the sample name The sample number is specified in field Sample description of the JCAMP file S AMPLE DESCRIPTION name number By default this option is not active only th
365. r the plane represented by the selected variable The 3 D Matrix Plot dialog allows you to select which sample resp variable to plot Four different layouts landscape bar map and contour are available Use Edit Options to change the layout of the plot Plot PCA Overview The option Plot PCA Overview gives you a quadruple plot of all the main results of a PCA namely the scores loadings influence and total residual variance as listed in the field Plot in the PCA Overview dialog where you specify how to plot the overview In the Components field in the same dialog you specify which components to use for the scores and loadings and for the influence plot respectively The Unscrambler Program Operation The Plot Menu e 177 Dialog PCA Overview PCA Overview x Cancel Plot Scores Loadings Influence Residual Variance Components Suggested no of components is 3 Scores and Loadings Vector 1 Vector 2 2 Influence P a Plot Variances and RMSEP There are three different ways to plot Variances and RMSEP against the number of components You can plot 1 X or Y variance 2 X and Y variance or X1 and X2 variance in the case of a three way PLS model 3 RMSE The last two options are only available if you have plotted regression results the first one is also available for PCA results with only X variance available You specify what to plot in the dialog Variances
366. ral Composite Design dialog 102 The Unscrambler Program Operation Index e 319 Extend a Centroid Design dialog 103 Extend a Fractional Factorial Design dialog 101 Extend a Full Factorial Design dialog 100 Extend a Lattice Design dialog 102 Extend an Axial Design dialog 103 Extend dialog 100 extract data from marked 288 from unmarked 288 F F1 button 38 file close 48 duplicate 80 duplicate as 2 D data table 81 duplicate as 3 D data table 81 duplicate as modified design 81 duplicate as non design 81 exit 96 export 71 export model 74 import 50 lock 83 new 45 new design 97 open 48 print 82 print lab report 82 print preview 83 print setup 83 properties 84 recent files list 95 save 49 save plot 49 unlock 84 file import plug in 65 file information 37 fill 120 Fill dialog 120 fill missing 120 Fill Missing dialog 120 121 filling missing values 31 find 127 flat ASCII file 53 flat wide ASCII file 53 54 fractional factorial design 98 extend 101 frame scaling 155 frozen calibration samples 259 Full Factorial Combination 114 115 full factorial design 98 extend 100 full screen 294 G gap 305 Gap Segment Derivatives dialog 229 230 gaussian filter 222 Gaussian Filter Smoothing dialog 222 general plot 184 general view 33 34 292 2D scatter plot 212 3D scatter plot 213 histogram plot 214 line plot 211 matrix plot 215 normal probability plot 215 Generate Design dialog 113 Get Centered Models dialog
367. rambler Program Operation The Task Menu e 277 In the Keep Out of Calculation field you enter any variables which you do not want to include in the analyses indicating ranges by hyphens and separating different ranges by commas for example 1 5 7 10 13 27 You can also choose to include interaction and square effects in the calculation by pressing the button I amp S which launches the Interaction and Square Effects Selection dialog At the bottom of the sheet a list displays all Estimated Effects Note You cannot have the same variable in both the X and Y variable set Task Analysis of Effects The Analysis of Effects dialog is used to compute effects and test their significance Analysis of Variance Analysis of Effects is only available for designed data tables The dialog contains three sheets Samples X variables Y variables as well as a Warnings field at the bottom Warnings In the bottom field of the dialog you define whether warnings should be issued by ticking or un ticking the Issue Warnings box Ticking it will enable the Warning Limits button which launches the Warning Limits dialog Analysis of Effects Dialog Samples Sheet All design and reference samples are used for Analysis of Effect hence the Sample Set field is frozen In the Keep Out of Calculation field you enter any samples which you do not want to include in the analyses indicating ranges by hyphens and separating different ranges by commas for e
368. ration Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results Statistics This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type Statistics Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results Classification This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type Classification Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results Prediction This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type Prediction Select the file you wish to view and press View to open a Viewer window where the results will be displayed Results MCR This command launches the Results dialog which gives you an overview of all your data and results files see detailed description p 289 The dialog is configured to display Unscrambler files of type MCR Select the file you wish to view and press View to open a Vi
369. re also given The plot can be generated from a response surface analysis an analysis of effects or a regression analysis using MLR The Analysis of Variance dialog will be slightly different depending on the number of levels studied for the design variables Analysis of Variance Dialog General Case In the Variables field of the Analysis of Variance dialog simply select the response variable Y variable of interest Double click on the plot preview in the dialog if you want the plot to be displayed on a larger resp smaller sub view Analysis of Variance Dialog Multiple Level Analysis of Effects In designs which contain variables with more than two levels i e Full Factorial designs you must select whether you want to plot an ANOVA table or Multiple Comparisons with the radio buttons in the Plot field in the Analysis of Variance dialog as seen in the figure below Then select the response variable Y variable and for Multiple Comparisons the X variable in the Variables field and decide whether you want the reference samples indicated in the multiple comparisons plot 202 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Analysis of Variance multilevel AoE OK Source OF SS Mean d Fvalu Cancel uly Help ANOVA Table r Plot Multiple Comparisons Variables w a 1 D Reference Sample Select j al Plot Effects This option lets you plot an over
370. re be evocative of the nature of the samples in the set The Unscrambler Program Operation The Modify Menu e 245 Dialog New Sample Set New Sample Set General Name Plant 2 K Cancel Help Interval valid range is 1 through 32 2 6 1 0 14 18 22 26 30 Select Special intervals Allsamples Every l4 samples counting from 2 Advanced fe In the Interval section enter the sample range Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10 13 27 Pressing the Select button next to this field launches an Editor the Select Samples dialog which allows you to select samples by clicking on them In the Special intervals field you can enter the set interval by using the currently selected samples in the Editor Every nth sample can also be entered into the set Specify the value of n with the spin buttons To enter the special set intervals into the Set Interval box press Update The Advanced field at the bottom allows you to select a Category Variable from a drop down list and automatically create one new Sample Set for each level of the Category Variable Note Samples with missing value of the Category Variable will not be included into any of the new Sample Sets New Variable Set Dialog Allows you to specify a new variable set or modify its properties In the General section give the new set a name in the Name field This is the name th
371. re measured in different units or have different scales For more details on specific cases of tables with missing data lookup chapter Fill Missing Difficult Situations p 121 File Recent Files List The list of recently opened files is displayed at the bottom of the File menu list just above the last menu option Exit The most recently opened file is listed first then the second most recent and so on The file type is displayed to the right of the file name as either d data table or r result file To re open a file from the list click on the desired file name or press lt Alt gt lt F gt and type the corresponding number Note The option is disabled if you are logged in as supervisor The Unscrambler Program Operation The File Menu e 95 By default the list contains up to 4 files You may change that number in the General sheet of the System Setup dialog as described p 91 File Exit Through this option you can quit The Unscrambler You are prompted to save any data files that have changed since the last time they were saved 96 e The Unscrambler Menu Options The Unscrambler Program Operation The Design Wizard File New Design Use this option if you wish to create a new design or to modify or extend an existing design The Design Wizard The Design Wizard can be accessed through two different commands If you chose File New Design the first dialog is Select Method to Use where you w
372. re to type in the same e mail address in the form as the one you registered as an Unscrambler licensee to receive a permanent activation key If you do not have Internet access Contact Camo and give your Camo consultant the Unscrambler License Server machine code that is displayed on screen Your consultant will provide you with an activation key in accordance to your network license The Unscrambler Acht Unscrambler machine code Obtain Unscrambler activation key Activate Currently The Unscrambler v9 6 is not activated Click on the Obtain button to request your Unscrambler activation key on line If you have no internet connection contact your Camo office and provide us with your Unscrambler machine code Cancel ddi Activating The Unscrambler License Server 12 Type in your Unscrambler License Server activation key into the relevant field of The Unscrambler Activation dialog then click the Activate button The Unscrambler License Server Activate Unscrambler License Server Machine Code 21D866C6 Unscrambler Licens eT _TDKSCZLRZTRKXBTCLGBNDX Activate Server Activation K Currently Unscrambler Licenses i ob Click on Obtain button to get the activation key Cancel For any queries contact support camo com or support camo no 13 A message confirms that the activation was successful The Unscrambler Program Operation Network Installation Multiple Us
373. reference values were included for comparison purposes when you made the predictions Note Deviations are used to identify outliers and nothing else Do not use them e g to express the prediction error In the Plot field you can also choose to include a table which contains the predicted value the deviation and the reference value of each sample in the prediction You cannot select several components when this option is selected Select the Y variable for which you want to look at the predictions and choose how many components to use in the prediction in the appropriate fields You should use the optimal number of components you found for the model based on validation Prediction Dialog Residuals Leverage Sheet From this dialog sheet you can specify a plot of prediction residuals or of the leverage for the prediction samples Those two elements participate in building the Deviations Use the plots to understand why some samples have larger deviations than the others 210 e The Unscrambler Menu Options The Unscrambler Program Operation nscrambler User Manual Camo Software AS Prediction Predicted Residuals Leverage TE Cancel Help CAMO r Plot Sample Residuals C Leverage Sample 2 E Obj256 Components Suggested no of components is 2 ba In the Plot field choose whether you want to plot Sample Residuals which show the X residuals of the samples after the specified number of components hav
374. rformed on the whole data table Modify Reverse Sample Order 3 D This option is disabled in 3 D tables with an OV layout since this would disturb the Primary Secondary structure of the samples Modify Reverse Variable Order This option reverses the order of variables in your data table from last to first All Variable Sets are automatically updated accordingly as with any other sorting option Note Using this option sorting is always performed on the whole data table Modify Reverse Variable Order 3 D This option is disabled in 3 D tables with an OV layout since this would disturb the Primary Secondary structure of the variables The Unscrambler Program Operation The Modify Menu e 237 Modify Swap 3 D Layout Choose Modify Swap 3 D Layout to re format a 3 D table from OV to O V or vice versa From OV to ON secondary variables become secondary samples From O V to OV secondary samples become secondary variables Sample variable names are also updated Primary variables and primary samples remain unchanged in either case The operation can be undone using Modify Undo Swap 3 D Layout or CTRL Z Modify Swap Samples amp Variables Choose Modify Swap Samples amp Variables to change the layout of a 3 D table in a flexible way The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z Here are all the choices to be found under the Modify Swap Samples amp Variables menu
375. riables correlate The command Plot Loadings launches the Loadings dialog where you can choose between three different sub view layouts for plotting and select various options type of plot variables PCs Loadings Dialog and Loading Weights Dialog The Loadings and Loading Weights dialogs offer three different layouts for plotting each available from its own sheet 1 General one plot window for PCA PCR and PLS two windows for three way PLS 2 2x2D Scatter two windows 3 4x2D Scatter four windows Loadings Loading Weights Dialog General Sheet On the General sheet which is active by default you specify either e one plot if you are viewing PCA PCR or PLS results The plot will be displayed either in the currently active sub view or in sub view 1 e two plots if you are viewing three way PLS results The plots will show results along the same couple of PCs for resp Primary and Secondary variables in the upper and lower sub view Note that if you choose only Y variables see the options below only one plot will be produced in the upper sub view You must decide whether you want a Line 2D Scatter or 3D Scatter plot in the field Plot Type and specify along which components you want the results in the Components field Use the default values 1 and 2 unless you have specific reasons for doing otherwise Finally make your choice between plotting either X Y or both X and Y variables in the Variables field
376. rial Design p 100 Design Wizard Extend a Lattice Design p 102 j N The Unscrambler Program Operation The Design Wizard e 97 Design Wizard Generate Design p 113 Design Wizard Last Checks p 118 Design Wizard Randomization Details General p 117 Design Wizard Select Design Type Design Wizard Select Existing Design p 100 Design Wizard Select Method To Use p 98 Design Wizard Select Method To Use This is where you can choose between starting from scratch modifying or extending an existing design by selecting the appropriate radio button Read the contents of the Information field at the bottom to get more details about the currently selected option Once you have made your choice click Next to access the next dialog Note Use Modify Existing Design only if the new design you want to create through the modification is of the same type as the existing one Select Method To Use Next Dialog e Ifyou have selected From scratch you access the dialog Select Design Type e If you have selected Modify existing design or Extend existing design you access the dialog Select Existing Design Design Wizard Select Design Type This dialog is accessed when you have chosen to build a design from scratch This is where you can build a completely new design In the Select Design Type dialog you can choose between three types of classical screening designs two types of cla
377. rial design Note All design variables must vary continuously between a low and a high level Box Behnken Design Optimization Design 3 6 An alternative to central composite designs when you have reason to believe that the optimum of your response is not in the corners of the experimental region and if you are not interested in extending a previous factorial design Note All design variables must vary continuously between a low and a high level D Optimal Non mixture You have only Process variables and some of these present Design Multi Linear Constraints Screening 2 12 Optimization 2 6 Mixture Design You have Mixture variables and possibly some Process variables in addition Some of the Mixture or some of the Process variables can be constrained s 3 15 or 32 Screening Axial Optimization 3 6 According to your specifications one of 4 designs can be used Simplex Lattice Design Simplex Centroid Design Axial Design D Optimal Mixture Design 1 A D Optimal Design will be used with Mixture variables if the experimental region is not a simplex in particular if some of your mixture variables present Multi Linear constraints Select Design Type Next Dialog It is easy to set up a design from scratch in The Unscrambler Once you have made your choice in the Select Design Type dialog and clicked Next you will enter a string of dialogs which tell you exactly what to do Simply fill in the necessary in
378. ross Validation the Setup button is disabled for this option Note It is dangerous to use leverage correction in connection with MLR if the rank of any submatrix is different from the rank of the whole matrix as the results cannot be trusted Run MLR with cross validation first and see whether you get a warning that this is the case If not you may use leverage correction on later models with the same data selection to save time If you choose the validation method Test Set press Setup to enter the Test Set Validation Setup dialog where you can further define your test set validation settings In the bottom field of the dialog you define whether warnings should be issued by ticking or unticking the Issue Warnings box Ticking it will enable the Warning Limits button which launches the Warning Limits dialog Regression Three Way PLS Dialog The Regression Three Way PLS dialog allows you to set up the details of your tri PLS nPLS regression model The Unscrambler Program Operation The Task Menu e 273 Dialog Regression Three Way PLS Regression Three Way PLS f Pri Xvars Sec X vars Y variables Sample Set All Samples BJ Define Keep Out of Calculation HESE Frozen Calibration Samples Match Samples in X and Y Data Tables By row numbers C By sample names Validation Method es f Cross Validation r EH e Test Set Setup Kl Maps 2 Iw Center Data r Seed The top
379. rs will be displayed in the field next to that button e g 3 7 12 17 21 27 When you are satisfied with your selection click OK to go back to the previous dialog else click Cancel The Unscrambler Program Operation The Modify Menu e 251 The Task Menu Task Statistics This dialog is used to compute the means standard deviations correlations and percentiles of measurements You specify the samples and variables you want to work with in the Statistics dialog which contains 2 sheets Samples and Variables as well as one general option You can also group samples according to levels of the variables in the data table Dialog Statistics Statistics E Samples Variables Sample Set Cancel All Samples 36 gt Define E Keep Out of Calculation fi 2 Subgroups D Variables 0 Groups Define M Calculate Cross Correlation rl General Options The box Calculate Cross Correlation is available regardless of which sheet you are on Tick this box if you want to calculate the cross correlation between the variables Once the statistics have been computed you may save the result file and later import the cross correlation matrix StatCorr using File Import Unscrambler Results Note Computing cross correlations may be time and memory consuming turn off this option if you do not need the results Statistics Dialog Samples Sheet Choose the Sample Set on which you want to base the calcu
380. rtainty Test values for the parameters can be imported from the result file These are also stored for PCA but the plotting options are not implemented 160 e The Unscrambler Menu Options The Unscrambler Program Operation he Unscrambler User Manual Camo Software AS When you have activated a stability plot you can click on individual Jack knifing points and a segment number will be displayed This information will help you in finding out which samples are causing the variation View Uncertainty Test Uncertainty Limits This command is available when a Regression Coefficients plot is currently active in a Regression result generated with the option Uncertainty Test The limits are given as twice the estimated standard deviation for the coefficients under ideal conditions this corresponds to approximately 95 confidence interval View Correlation Loadings EI When a PCA PLS or PCR analysis has been performed and a loading plot is displayed on your screen use View Correlation Loadings to take into account the amount of explained variance in your interpretation The correlation loading plot will help you discover the structure in the data more clearly Below is a comparison between the ordinary Loadings plot and the Correlation Loadings plot The importance of individual variables is visualized more clearly in the correlation loading plot compared to the standard loading plot As an example in the figures below the imp
381. s Plots can be inserted into other programs either as a Bitmap or as a meta file Picture You can choose one of those options in the Clipboard format field If the plot has many objects many samples and or variables using symbols layout etc it will occupy less space in the other programs if you choose the Bitmap option For plots with few objects it may be the other way around In the Miscellaneous field you have 4 choices e Large data matrices more than 5000 elements may be slow to plot By checking the box Disable browser info for large plots the plotting time is reduced The disadvantage is that you can no longer click at a data point to get information regarding that point e By checking the box Remember plot settings you enable easy customization of the result plots you want to display when opening a previously saved results file The plots appearing in the Viewer when you open a model file will be the same ones that were on screen the last time you saved the model Un check this box to disable customization e Check the box Activate overview plot when entering viewer if you want the overview plots to appear when you click View in the progress dialog after an analysis If this choice is not selected the Viewer will be empty and you have to select a plot from the Plot menu unless you have checked the previous box Remember plot settings in which case the current choice is grayed out e Markers labels that are blank or contai
382. s the plot will display RMSEC e If you select Validation samples the plot will display RMSEP In the field Variables you select which Y variables you want to plot To select All variables click this button and to enter an Editor where you can select variables click Select You can decide whether you want the variance to be plotted for calibration or validation samples respectively or both by ticking the appropriate box es in the Samples field Calibration variance measures the fit of the model to the calibration data The validation variance is an estimate of the fit of the model to new data not present in the calibration and is based on the validation Plot both measures to obtain full information about model fit You can later change that choice by using View Source The dialog displays a preview screen Double click on this to change the size of the plot 180 e The Unscrambler Menu Options The Unscrambler Program Operation Plot Sample Outliers This quadruple plot is useful for visualizing sample outliers Both validation and calibration samples are plotted in the Sample Variance and Score plots while the influence plot contains only calibration samples In the Sample Outliers dialog you must select the components for the Scores plot and Influence plot in the Components field Start with the first few components or the component number where the prediction error starts to increase after a minimum Plot Scores and Loadin
383. s 0 0 cceecccecessecseeesessesseeseceeeceeeeeeaeeeesaeeae eeesseseaeeeeeeeeeeneeneeeaes 170 Views Source EE EE 170 View Source Y Variables ccecesceesceseesseeeseeecseesseeseeeeeeeeececaeeaeeceeesseseeceneceeeeeneeeeaees 170 View Source Explained Variance pairere esaia r e Ea E 170 View Source Residual Vartance ceceeeceeeeeenaeeaees 171 View SOUTCE TEE 171 View Source Studentized Residuals cccceceeceesceeeceseeeceecenecneeeseceseeeeeeeeceeeeceeeecneeeeenas 171 Mew Raw RE 171 View MCR Message List niriana ernn e ieee E E E S 171 Miew Outlier Lis EE 172 View LOOM DAIS iss secin arena tates ea a E Rr E a EE SE o RE TEE N EENE ENEs 173 NASW lt S tatus EE 173 WME ae Meng E AT E E EES 175 The Unscrambler Program Operation Contents e vii Plot EE 175 Plot Date aere SEENEN OPETE EEA EEEE OERE EEEE EEEN EEEE 175 Plot 3D SCAtter Egeter ENEE r tosses AE ETEEN E EE E E E E EiS 175 Plot N rmal Probability oc icsssctesissvses CEO Ce dree EIERE AEE A EE 176 POU Histo EE eege seas a e dE teavecteavens eae 176 lot Matti xe eisene eve ices eis chew det sieve oem 177 Plot Meter eebe EE eege Ee Eer 177 Plot PEA EE 177 Plot Variances and RMSEP AAA 178 Plot Sample Outhers seccssssesccssseevtetescciensoesssctbeces ossevecoesies sane rE EE ns AEREE TESE SSES 181 Plot Scores and Loading Seasin erinrar eeen i EEEE EE AEEA EEEE EEEE EEKEUR 181 Plot Scores and Loading Weights AAA 182 Rio EO e REEE
384. s 161 185 2 x 2D scatter plot 185 4 x 2D scatter plot 186 general 185 Loadings dialog 185 Locate UDT Dialog 234 235 lock 83 locked cell 31 login dialog 26 Login dialog 26 Lotus files 55 Lower Bound 89 lower quartile 306 main window 29 mark 146 evenly distributed samples only 147 one by one 148 outliers 146 reverse marking 148 samples 31 test samples 146 unmark all 148 variables 31 with rectangle 148 Matlab files export 72 import 61 import 3 D 69 matrix plot 177 matrix 3 D plot 177 matrix plot general view 215 Matrix Plot dialog 177 maximum normalization 222 223 MCR 267 MCR dialog 267 MCR message list 171 MCR overview 190 MCR Overview dialog 190 mean normalization 222 223 median filter 221 Median Filter Smoothing dialog 221 memory space 1 menu 29 File Convert Vector to Data Table 80 File Import 3 D 66 View Rotate 157 Menu File Convert Vector to Data Table 80 File Import 3 D 66 menu bar 29 meta file 92 93 min max scaling 155 missing data fill in the Editor 31 MixSum 306 Mixture design D Optimal Mixture 115 Mixture Design 89 D Optimal Mixture 114 mixture sum 306 Mixture Sum 89 107 108 129 131 306 mixture variable 307 MLR 268 270 modes 307 modify compute 217 compute general 217 edit set 246 edit set 3 D data 246 layout 245 properties 244 redo 243 reverse sample order 238 reverse variable order 239 shift variables 238 sort samples 235 sort samples by sets 236 sort variab
385. s Menu 289 The Window Menu 292 The Help Menu 297 In addition the Design Wizard launched by File Menu options File New Design and File Duplicate As Modified Design is described in a chapter of its own starting p 97 The File Menu File New D Ctrl N You use this option if you want to create a new data table by typing the data directly into the Editor or by importing foreign files You enter a dialog called New Data Table where you have a choice between three layouts for your data The Unscrambler Program Operation The File Menu e 45 Camo Software AS The Unscrambler User Manual Dialog New Data Table accessed from menu File New New Data Table Ea m Format Plain 2 D data table Cancel 3 D data table layout O objects x variables x variables Help C 3 D data table layout DEM objects x objects x variables L m Size Primary variables i Secondary variables 12 Samples IEN SS Secondary samples ah l ap Choose a layout see below for more details and fill in desired number of variables and samples Upon clicking OK you enter an Editor window containing an empty data table with the specified size Fill in the row names column names and values for your data Plain 2 D data table Specify the size of the table by filling in the fields Variables and Samples Type in the desired number or use the spin button to tune the value up and d
386. s available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices Properties Dialog Model Sheet The Model sheet is only available for result files This is where you set the number of components you want used as default when the model is used for prediction or classification This is necessary if the number of components you want to use is different from the optimal number of components suggested by The Unscrambler You can override this default in the prediction and classification dialog 88 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Properties Model Sheet Tutor c bascorr signif Properties General Notes Transformations Model This model can be used in Classification and Prediction analyses The number of principal components PCs to use in these analyses can be specified here Number of PCs in Classification amp Prediction E ot Properties Dialog Constraints Sheet This sheet only applies to data tables In this page find a summary of the settings for your D Optimal or Mixture design e Recapitulation of the multi linear constraints e MixSum value e For non designed tables with mixture variables the MixSum value can be changed in this page Values which do not respect the lower and upper bounds of your mixture variables will be rejected The Unscrambler Program Operation The File Menu e 89 The Unscrambler User Manual Dialog Data Tab
387. s between the last selection and the present selection When you make a selection you always mark either samples or variables i e you either select some variables for all samples or some samples for all variables You can also mark the whole matrix but the selection is still sample or variable oriented You can see whether you are marking samples or variables by looking at the shape of the mouse pointer as you make the selection The shape of the mouse pointer when marking samples and variables respectively Mark Samples Mark Variables The Editor Screen Layout If the data table is larger than the screen you can scroll the Editor Information about the active cell is displayed in The Unscrambler s status bar Variable names are displayed in black if the variable is continuous and in blue if it is a category variable Locked cells e g design variables are grayed out to show that they cannot be edited The Editor Filling Missing Values If a data table contains some missing values those may be filled with estimated values that take into account the underlying data structure There are two ways of doing this e One table at a time use menu option Edit Fill Missing when your table is displayed in the Editor This allows you to select the scope of the replacement it need not apply to the whole data table e Automatically fill missing values for all data tables use menu option File System Se
388. s cannot be placed simultaneously in the list of Selected Sets Click Define if you need to define new sets this will lead you to the Set Editor dialog Modify Sort Variables by Sets When using this option all variables belonging to Variable Set 1 are displayed in the first rows of the data table followed by variables from Variable Set 2 etc The sorting order is defined by the alphabetic order in which the Sets are displayed as a list in the Set Editor Variables that do not belong to any set are displayed at the end of the table If the variable sets currently defined in your data table are overlapping the command opens the Select Variable Sets To Sort dialog where you can choose which sets must be taken into account for the sorting operation Note1 This option is only available when at least one Variable Set is defined Note2 You cannot sort variables with locked cells e g designed data Modify Sort Variables by Sets can be undone redone with the corresponding options of the Modify menu Modify Sort Variables by Sets 3 D Sorting is disabled in 3 D tables with an ON layout since this would disturb the Primary Secondary structure of the variables Modify Shift Variables You can shift variables up or down in the data table This is often used for time related measurements e g process data Shifting variables will generate time lags between the various measurements It is a column oriented transformat
389. s in the data table are known to be pure constituent spectra or assu med to be an approximation of those Type in the sample range directly or click Select to select relevant rows from an Editor window The corresponding sample numbers will be updated in the Initial Guess Samples field of the Samples sheet Note Variables resp samples defined as Initial Guess cannot be part of the model data In case of overlap an error message appears when you launch the computations as shown hereafter The Unscrambler xj A Initial Guess Samples Yariables should not overlap with Sample Variable Sets To fix the problem remove the offending variables samples from the model data by typing in their range in the Keep Out of Calculation field MCR Dialog Constraints Setting This group of options specifies the constraints to be taken into account in the model The following choices are available e Non negative Concentrations Variables is selected by default This is the most common case e Non negative Spectra Samples is selected by default In many applications the instrumental measurements e g UV absorbance or fluorescence intensity can be assumed to be non negative e Closure is not selected by default Choose this option if you have reason to assume that the sum of the concentrations is constant over all samples e Unimodality is not selected by default Choose this option if you want to allow for only one maximum per profile
390. s of observed or computed values in such a way that their distribution can be visually compared to a normal distribution The observed values are used as abscissa and the ordinate displays the corresponding percentiles on a special scale Thus if the values are approximately normally distributed around zero the points will appear close to a straight line going through 0 50 A normal probability plot can be used to check the normality of the residuals they should be normal outliers will stick out and to visually detect significant effects in screening designs with few residual degrees of freedom ON In The Unscrambler three way data structure formed of two Object modes and one Variable mode A 3 D data table with layout ON is displayed in the Editor as a flat unfolded table with as many rows as Primary samples times Secondary samples and as many columns as Variables The Unscrambler Program Operation Glossary of Terms e 307 Offset See Intercept ov In The Unscrambler three way data structure formed of one Object mode and two Variable modes A 3 D data table with layout OV is displayed in the Editor as a flat unfolded table with as many rows as Objects samples and as many columns as Primary variables times Secondary variables Passified When you apply the Passify weighting option to a variable it becomes Passified This means that it loses all influence on the model but it is not removed from
391. s of the current model The residuals are the difference between the observed values and the values predicted by the model The size of the residuals tells you about the amount of error in your model A large residual shows that the sample is not well described by the model View Source Studentized Residuals Displays the studentized residuals Studentized residuals are scaled and corrected for leverage The scaling makes the interpretation independent of the measurement units and the leverage correction increases the residuals for samples far from the model center View Raw Data When you select this option from a result Viewer an Editor containing the raw data used to make the model you are looking at is displayed This Editor and the Viewer are linked so marked samples or variables in the Viewer are marked in the Editor and vice versa This is handy when you cannot find a sample or variable in the plot Mark it in the Editor and you quickly see it in the plot As the Sample Statistics and Variable Statistics tables the Raw Data table behaves as a slave It cannot be edited or plotted If you want to freely manipulate the values in that table you have to drag and drop its contents to an empty spot of the workplace View MCR Message List This option is only available when viewing MCR results It allows you to display the list of system generated messages related to your MCR results The MCR message list appears in a sep
392. s should normally result in a solution with a lower SOD if not try again a few times before giving up Go on running new analyses with increasing Number of Iterations until you no get no more improvement on the resulting SOD Pick the clustering variable with the lowest SOD as your final result Note Since the first step of the K Means algorithm is based on the random distribution of the samples into k different clusters there is a good possibility that the final clustering solution will not be exactly the same for every instance for a fairly large sample data set Task PCA In the Principal Component Analysis dialog you define the data on which the model will be based the maximum complexity of the model and how to validate the model The dialog contains two sheets Samples and Variables and two general groups of options Validation Method and some miscellaneous options Dialog Principal Component Analysis Principal Component Analysis ES Samples Variables Sample Set Cancel Wl Samples Help ill Keep Out of Calculation Select Frozen Calibration Samples fo Select Validation Method C Leverage Correction Ze Cross Validation Setup 3 Ca HPCs C Test Set Setup Model Size Full v NumpPts 3 IV Center Data J Add Start Noise IV Issue Wamings Warning Limits 258 e The Unscrambler Menu Options The Unscrambler Program Operation
393. s you have This option is only available when you start the Editor from a result plot in the Viewer using View Raw Data View Level Indices Replaces the level names with level indices 0 1 2 in all levelled variables Category or Design variables This option is only available when the currently active table contains at least one levelled variable View Standard Sample Sequence This option only applies to designed data If you activate the option samples in the Editor are shown according to the standard design order Use this option to see the logic of the design View Experiment Sample Sequence This option only applies to designed data Through this option samples in the Editor are shown in randomized order i e in the same sequence as the experiments are supposed to be performed View Graphical Lets you view the selected data of a Viewer in graphical mode Use this command to return to a graphical view after examining your results as a numerical table View Numerical View Numerical Through this option you may display results plotted in a Viewer as a numerical table You can copy that data table to the Clipboard and paste it into an Editor or another software package Restore the plot using View Graphical or Ctrl G View Autoscale H or Ctrl H 154 e The Unscrambler Menu Options The Unscrambler Program Operation Scales the plot so that all data points are shown within the Viewer window This
394. sample resulting in 7 new rows Help Edit Insert Category Variable Here you may insert a category variable which is useful during interpretation of plots and to create groups on which to calculate statistics Note You cannot insert a Category variable into a 3 D table with layout OV This option starts the Category Variable Wizard which consists of the following dialogs Category Variable Wizard Enter Variable Name and Choose Method In the field Category variable name you enter the name you want to give the variable in the data table The name will be written in blue in the Editor to indicate that it is a category variable Then you can choose between specifying the levels of the variable manually or basing the levels on a collection of sample sets The levels are the discrete values that are valid for that variable Select the appropriate radio button and press Next to go to the next dialog which will depend on your choice of method e If you selected manually the next dialog is Specify Levels e If you selected collection of sample sets the next dialog is Select Sets Category Variable Wizard Specify Levels This dialog lets you specify your levels manually To create a new level 1 Type in the name of the new level in the empty field to the left of the Add button 2 Press Add to validate the new level To change a level name double click on that level on the list and type in a new nam
395. se No the setup will close To install The Unscrambler you must accept this agreement No 22 Select a destination folder for the The Unscrambler Program files InstallShield Wizard Choose Destination Location Select folder where Setup will install files Setup will install The Unscrambler in the following folder To install to this folder click Next To install to a different folder click Browse and select another folder m Destination Folder C Program Files T he Unscramblers Browse 23 Select a destination folder for the The Unscrambler DATA files O H m O Software A wn The Unscrambler Program Operation Network Installation Multiple User Licenses e 17 Camo Software AS The Unscrambler User Manual InstallShield Wizard Choose Data Files Location Setup will install the The Unscrambler data files in the following directory To install to this directory click Next To install to a different directory click Browse and select another directory If this installation is going to be accessed by several users in a network select a directory on a file server that all users can reach Also make sure that all users have write access to this Destination Folder C My Documents The Unscrambler DATA 24 The program is being installed InstallShield Wizard Setup Status The Unscrambler Setup is performing the requested operations C Program Files The
396. se details 204 Effects dialog 203 EMSC 225 EMSC dialog 225 227 end point 305 Enter Multi Linear Constraints dialog 109 Enter Variable Name and Choose Method dialog 124 estimated concentrations 191 Estimated Concentrations dialog 191 estimated spectra 191 Estimated Spectra dialog 191 Evenly Distributed Samples dialog 147 exact frame scaling 156 exact min max scaling 156 Exact Scaling dialog 156 Excel files import 55 import 3 D 68 exit 96 experiment sample sequence 154 Experimental error 114 experimental region not a Simplex 115 Simplex 115 Experimental Region not a Simplex 114 Simplex 107 108 114 export 71 ASCII files 72 ASCII MOD 75 ASCI MOD format 41 data 40 JCAMP DX files 72 Matlab files 72 model 74 MVACDF files 74 NSAS 76 NSAS transformations 77 plots 42 result matrix 41 supported formats 71 74 to spreadsheet 40 to word processor 40 Tracker model 79 U5 data 71 U5 model 76 Unscrambler ASCII data 71 useful tips 40 Vision 79 Export 71 Export ASCII dialog 72 Export ASCH MOD dialog 75 Export dialog 71 Export Matlab dialog 72 Export Tracker Model dialog 79 Export U5 Model dialog 76 Export Unscrambler 5 dialog 71 Export Unscrambler ASCII dialog 71 extend axial design 103 central composite design 102 centroid design 103 data table 125 existing design 98 fractional factorial design 101 full factorial design 100 lattice design 102 to central composite design 100 101 to full factorial design 101 Extend a Cent
397. se into The Unscrambler s Editor Since there are many possible database platforms and the data structure may be complex you need to go through several dialogs in order to specify the import 1 Data Link Properties dialog 2 SQL Query dialog If you use File Import Database several times without logging off from The Unscrambler the following message box will appear The Unscrambler a xj AN Warning You have established a database connection before Do you want to re establish the same connection If you click Cancel the Data Link Properties dialog is bypassed keeping the same connection details as previously and you access the directly Click No in order to access the Data Link Properties dialog and specify new connection details The Cancel button cancels the File Import Database command Note The Data Link Properties dialog is a standard Windows dialog Depending on your local language setup this dialog may be displayed in another language than English The name of the dialog will be different the fields will have a different text but the layout and meaning of all fields will be the same as described hereafter The next two sections describe the standard stages you need to go through in order to establish a connection from The Unscrambler to a database For addidional information click Help this will start the Microsoft help system related to the current sheet in the Data Link Properties dialog Data Lin
398. segment are listed in the field Samples not found in any segment Select the Number of Segments you want in the cross validation or how many samples there should be in each segment in the appropriate fields These two options are inter dependent any change you make in one is automatically reflected in the other The Samples in Current Segment box shows the sample numbers of the selected segment in the segment list The box is opened for changes only when manual cross validation is selected Test Set Validation Setup Dialog The Setup button in the Validation field in a dialog will also be enabled if you choose the validation method Test Set Press Setup and you enter the Test Set Validation Setup dialog You can select the test set samples manually or by group or let The Unscrambler pick them at random by choosing the appropriate radio button If you choose Manual Selection enter the test set samples in the appropriate box or press Select to enter the Set Editor where you can easily choose the samples you want If the data table contains category variables or if it is a designed data table you may use Group Selection i e a level as a test set Select which variable you want to group by and whether the test set should be equal to or differ from the level in the box to the right If you select Random selection you only need to specify how many samples you want in the test set The Unscrambler picks these at random from t
399. select which X variables you want to vary with the combo boxes Use the Configure to invoke the Configure dialog where you can set up which X variables vary along the axes and which constant values the others should take Plot Response Surface Configure Dialog In the Configure dialog you can change the ranges of variation for the varying variables and the fixed values for the non varying variables Dialog Configure X variables Axis Min Max ok 1 Lamp tits f RR Eso x Ced 3 fue zl Variable Level Value Milk Varied Flour B Varied eo 22 708 salt D Varied ad Help Variable eo 22 70833 Choose which variables you want on axis 1 and axis 2 and axis 3 for a mixture plot of the response surface plot from the drop down lists in the X variables field The Unscrambler Program Operation The Plot Menu e 201 The Min and Max boxes enable you to select the ranges of variation for the varied X variables these boxes are disabled when making a response surface plot from designed data The other X variables are by default set to their mean value You can change this by selecting the variable from the list in the Variable field and using the spin buttons on level and value If a category variable is chosen the spin button changes to a list of the available levels Plot Response Surface Overview This is a quadruple plot of some of the main results of a response surface analysis n
400. sheet dialog which pops up after choosing File Import Excel or Lotus The Unscrambler Program Operation Useful Tips for Import Export e 41 Camo Software AS The Unscrambler User Manual The Import from Worksheet dialog with range names Sheet name Sheet1 K Range names Sheet range Rows Columns Cancel Data Values Feu 40 2 Help Sample names s amples aaa 40 1 Variable names Variables B1 Del 1 2 ddi Note The source file containing the spreadsheet from which you want to import should not be opened by another application e g Excel while you are performing the import Transferring Unscrambler plots to Another Application You can use copy and paste to transfer a plot from The Unscrambler Viewer to another program Click on the plot you want to transfer and select Edit Copy or select Edit Copy All if you want to copy all plots together Then go to the other application and select Edit Paste The plot can be transferred either as a bitmap or as a meta file picture The latter option gives the best quality but may result in very large files in Word especially if the plot has many objects many samples and or variables using symbols layout etc You select the appropriate option in the Viewer sheet of the File System Setup dialog Another possibility is to save the plot using File Save Plot or right click on the plot and select Save Plot This allows you to choose among a few differe
401. signed data table Restrictions 109 non negativity 307 normal distribution checking 307 normal probability plot 176 307 general view 215 Normal Probability Plot dialog 176 normalization 222 area 222 223 maximum 222 223 mean 222 223 peak 222 224 unit vector 222 223 Normalize dialog 222 223 Norris Gap Derivatives dialog 229 N plot 176 nPLS analysis 273 prediction 284 NSAS export model 76 file instrument parameters 76 import data files 61 transformations 77 NSAS Export dialog 76 numeric qualifier 53 54 numerical 154 table 37 O 02V 307 offset 305 308 Open File dialog 48 options 134 general sheet 134 sample grouping sheet 135 Options dialog 134 other applications 39 outlier 193 list 172 Outlier List dialog 172 outlier marking 146 outlier warnings 306 OV2 308 overview plot 33 34 overview plots 92 93 OW 172 P passified 308 passify 308 paste 119 PC 309 PCA 258 overview 177 PCA dialog 258 PCA Overview dialog 177 PCR 268 269 peak normalization 222 224 percentile 306 308 313 Plackett Burman designs 98 99 planes 307 plot 2D scatter general view 212 3D scatter general view 213 histogram general view 214 line general view 211 matrix general view 215 normal probability general view 215 The Unscrambler Program Operation Index e 323 plot 2D scatter 175 3 D 157 3D scatter 175 analysis of variance 202 classification 208 effects 203 estimated concentratio
402. sion coefficients in the Y variable field You must also select the number of components for the computation of the coefficients in the Components field The regression coefficients are summarized for all components up to and including the chosen number For the loadings all components up to and including the chosen number are plotted in one sub view Plot Important Variables Three Way PLS Results In the Important Variables dialog you may configure your double plot of Regression Coefficients and Loading Weights The Unscrambler Program Operation The Plot Menu e 195 Dialog Important Variables Three Way PLS Results Important Variables OK Cancel pie Help Plot Variable Mode Regression Coefficients BW e Primary Loading Weights C Secondary Y variable 1 ES Severity EN Components Suggested no of components is 4 e Inthe X variable mode field choose between Primary and Secondary variables This will only affect the plot of Loading Weights the Regression Coefficients will be plotted as a matrix showing X1 and X2 variables together e Inthe Y variable field select the Y variable for which you want to display the regression coefficients e Inthe Components field select the number of components taken into account for the computation of the regression coefficients The regression coefficients are summarized for all components up to and including the chosen number For the loading weights all components
403. sis It includes two options 1 Stability Plot 2 Uncertainty Limits View Uncertainty Test Stability Plot l Use this feature in order to visualize the stability of the model by showing perturbations around the main model This will also indicate whether your uncertainty in Y is larger than in X This function is accessed from the viewer menu View Uncertainty Test Stability Plot or from the specific icon shown above This function can be applied when you are in any of the following 2 D scatter plots Score Loadings or X loading Weights and Y loadings They are available for PCA PCR and PLS models Score plot showing Jack knife segments Stability Plot RESULT2 X expl 27 31 Y expl 91 3 The circle in the score plot corresponds to the rotated score for a given sample for the segment where this sample was omitted from the model It is useful to interpret the influence of this sample see figure above X loading Weights and Y loadings Plot showing Stability results 0 6 0 3 0 0 3 0 6 RESULT2 X expl 27 31 Y expl 91 3 The spread around each variable gives information about the stability of the model and each particular variable Variables close to the origin often have larger spreads which is an indication that they are not significant When a variable has some extreme perturbation it could be due to the fact that this variable has a skewed distribution The individual Unce
404. site www camo com TheUnscrambler Appendices Useful Tips for Import Export The easiest way to share information between applications is to copy data tables and plots to the clipboard and paste them from the clipboard into other applications Other ways to do it are e Export data models or result matrices e Import data e Transfer data from spreadsheets e Transfer plots Exporting Data from The Unscrambler to Other Programs Sometimes you want to use data tables from The Unscrambler in your reports The contents of the data table may be raw data or model results e g regression coefficients The procedures for exporting to a word processor and a spreadsheet are different Transferring Your Data to a Word Processor Mark the data you want to transfer and select Edit Copy Go to the word processor and select Edit Paste The data is now transferred with sample and variable names Select the data and convert them to a table e g Table Convert text to table in Word and format this table as you wish to make it look nice Transferring Your Data to a Spreadsheet Export the data to an ASCII file select File Export and set the file type to ASCII The ASCII file should be exported as a block file with no name qualifiers Go to your spreadsheet and open the ASCII file In Excel a wizard will guide you through the process In step 2 you should change the delimiter to the one you chose in the Export dialog which
405. small correlation One model is made per Y variable Next you must specify which Sample and Variable Sets to use on the Samples X and Y Variables sheets respectively The Samples sheet is identical for all the methods while the X and Y Variables sheets and the common options at the bottom of the dialog differ slightly for PLS or PCR on one side and MLR on the other side Regression Dialog Samples Sheet In the Sample Set field select the sample set on which to perform the calculations Use Keep Out of Calculation if you wish to leave out some of the samples Finally define the Frozen Calibration Samples in the corresponding field For more details you may lookup the description of a similar sheet in PCA Dialog Samples Sheet p 259 Regression Dialog X Variables Sheet and Y Variables Sheet PLS and PCR On both sheets choose which variables to use in your calculation in the Variable Set field You can either choose an existing Variable Set from the drop down list or define a new in the Set Editor which you enter by pressing Define In the Keep Out of Calculation field you enter any variables which you do not want to include in the analysis indicating ranges by hyphens and separating different ranges by commas for example 1 5 7 10 13 27 On the X variables sheet you can also choose to include interaction and square effects in the calculation by pressing the button I amp S which launches the Interaction and Square Effects Se
406. snsnevees 290 Results Stausti cs oz seiten egen ee gege E A AAE nates estes 291 Results Classification ccceccecccscesseeeeeesecsenseeseceseceeeeeneeaececeseese caeesaeeaaeeaeeaeeeeeeaeesaeseaeaees 291 elteren sas 222 EE EE GE ee eg 291 Results MER uge NEES dE CECENE Ey EADAE ENEE EE dE 291 Results MSC EMS Cig erties TE A E ER SEEEN E E 291 Results RECKEN 292 LE e EH 292 THE Window EE 292 Window Copy TO EE 292 Window GO TO EE 293 Window Pull Sere E 294 Window Cascade cecceccssssssescenceeseeceeceeeecseesecseceseceeecees ceceeaeeecsaecaeseseceeeeeeseaeeeeeeeeneeaees 294 Window Tile Horizontally eee cnor na 294 Window Tile Nergcallwy ce ccesessesceceseceeeeecseseeesecaseeessesaeseeeesaesaseesseseseeseeseesenaaens 294 Window Close Current EE 294 Window Close TE 295 Window Identification eececescesesecceneaceeeeeeeceeececsaessessaeesceeceaeceeeeneeeaeeeesaesaeceaeeneneeeas 295 Window Warning Lists icsesscssisccsesscessesstasssbcaststtoesssssdacasieieerss se sasteepenss eE E a 295 Shift to Other Window 3 297 Th Help Menu igure en EAE EAR TEE EE ENEE dree RETE 297 Help The Unscrambler Help Topics cscsscscecsccsssesensosesssosevssestecceesenssensevssessbn sessavoseseseosers 297 Help Bt sii scssceseisstessvescssasecsiadasssesvavanedd EEEE r AEAEE AEREA I EEIE APEEP TE EREEREER 298 Help Tutorial Exercises oo ce ecesceeceeeee ceeeeseeesseeesecsesseeaecaseaeeeesesaeeesaceeseessaeeeseessaeeeets
407. ssical optimization designs and two additional types of design by selecting the appropriate radio button and clicking Next Different Types of Experimental Design The designs with their fields of application and the allowed number of design variables are listed below Available types of experimental design Field of Use Full Factorial Design Screening Design 2 6 a You have few design variables 2 to 6 and want to study all main and interaction effects independently from each other b One of your design variables has more than 2 levels Fractional Factorial Screening Design 3 15 Design Most useful in the early stages of a research project when the number of design variables may be large and the goal is to find 98 e The Unscrambler Menu Options The Unscrambler Program Operation out with a small number of experiments which variables should be investigated further Note This design only allows 2 levels for each design variable Plackett Burman Screening Design 4 32 Design An alternative to fractional factorial designs in the early stages where you study main effects only The number of experiments is minimized never more than number of design variables 4 Note This design only allows 2 levels for each design variable Central Composite Optimization Design 2 6 Design Finds the optimal settings of your design variables by adding a few more experiments to a full facto
408. st 2nd 3rd or the 4th derivative of the samples from the drop down list You may then select the smoothing points Note that a larger range will give a smoother shape to the sample but you may lose some information There are two ways of selecting the smoothing points e Either type in the Number of smoothing points it is always an odd number 2n 1 and the smoothing segment will be symmetrically distributed on both sides of each curve point n left side points N right side points e Or choose the Number of left side points and right side points separately Enter how many variables data points should be taken into account on the left and right side of each cell in the Averaging sub field Finally enter the Polynomial order i e the order of the polynomial to be fitted A polynomial order of 2 means that a second degree equation will be used to fit the data points A higher number means a more flexible polynomial i e a more precise differentiation Note that after the operation is completed the data will be slightly truncated at both ends If p is the number of left side points and q the number of right side points in the smoothing segment the first p and the last q variables in the smoothed variable set will be set to zero This is because there are not enough points to the left resp right of these variables to compute the smoothing function Modify Transform Baseline Baseline offset and Linear baseline correction ar
409. stance and the best result the one now displayed in the data table after 50 iterations was a sum of distances of 80 7654 then the clustering variable is called Euclidean_SOD 80 7654 Clustering Dialog Samples Sheet Choose the Sample Set on which you want to base the calculations from the Sample Set drop down list The K means clustering will be performed for the entire chosen sample set If you want to define a new set click Define to launch the Set Editor dialog If you want to exclude some of the samples from the analysis type in their numbers in the field Keep Out of Calculation or click Select to choose them directly from an Editor the Select Samples dialog Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10 13 27 Clustering Dialog Variables Sheet On the Variables sheet you decide on which Variable Set to perform the calculations in the Variable Set field Only continuous variables may be chosen If there are any category variables in your selection they will be automatically excluded from the analysis To define new variable sets click Define to go to the Set Editor If you want to exclude some of the variables from the analysis type in their numbers in the field Keep Out of Calculation or click Select to choose them directly from an Editor Editor the Select Variables dialog Ranges are indicated by hyphens and different ranges are separated by commas for example 1 5 7 10
410. t GRAMS is ccaes estscescscnsstated an a bebatisey bunaleltsvedeezs eebe Dese 61 SU e Matlab E 61 File Import MV ACD oe sesssevsesueetev cavestevssins dette aa E E E EESE 62 File Import CLASS PA amp SpectrOn 20 ce cceeceeesceee ccneesaeescnecaeeesaesaeeaeeaeeesseeeseaees 63 File Import Indico se secncsssceecteseseietacsesccecncsnstestencadesensteaseceteseasessetentesnsobecsdecsbessosevbeebesnsutes 64 File Import BEES EE 64 File Import User Defined Import UDI sssesesesesseserererrrrrisrsrseseresrsresererstrererererrerere enteres 65 Fil Import E TE 66 File Import 3 D Unscrambler Results 67 Fil Import 3 D ASCM EE 68 Files Import 32D EXCel enserrer rE EE EEE E E EEEE a EE Ta 68 File Import 3 D JCAMP DX persion osoin aE AE NANE EEE 69 iv e Contents The Unscrambler Program Operation Pil hnport 32D Matlab wc nnn canis ain ean Eder ee 69 Prl JanipOrt 3 D BSD E 69 Pile UDI e A obunasdevvcubensbsepagnsbssteasesnsutonsiriesetes 69 File UDI Register New UDI sessssccesdncevsccsesssssesta inde pssensecanscvesssovsseazseersetesteaznalsvveteneaecessene te 70 File UDE Unregister EIERE graegetie dree deet oo ested ei E EE 71 Piles EX POT EE 71 SIE Elle E EE 74 File Export Model ASCTI MOD 00 ceeeesecseeseeeesecaeeeetaecaeeessaeseseeesanaceeseesaeeeenees 75 File Export Model US Model 76 File Export Model NSA ysicsssssghdsessssstztelssscaeeetssiescavscevs neccst
411. t xj Sheet name pure components DI IV First column is sample names V First row is variable names Cancel Range names Sheet range Rows Columns Help Data user range gt E 2D37 36 3 cher fi kuser range gt DI A2437 36 1 Gm Les z v0 dE The check boxes right under Sheet name take care of the most common cases where First column contains all sample names or First row contains all variable names Un tick them if not relevant The rest of the dialog box lets you specify where to find the data to be imported It takes advantage of the Name function in Excel which lets you allocate a name to a certain area or group of cells in Excel use Insert Name Define to define a range and give it a name All ranges that have been defined with names in the selected Excel sheet are listed under Range names The sheet range is updated automatically if you select a range name You can also enter the range manually specifying the upper left cell and the lower right cell eg B2 H34 All cells lying within this rectangle are then imported Select the appropriate ranges as described above for the data values as well as for the sample and variable names if relevant Note You must clear the sample and variable names fields if you do not want to import such names The Unscrambler Program Operation The File Menu e 55 Camo Software AS The Unscrambler User Manual File Import Database You may import data from a databa
412. t Noise Sr Ze Za ba Issue Warnings Number of PCs In the Num PCs field you define the maximum number of Principal Components to compute Choosing many components will give you a model which explains more of the variations in the data but which also takes more time to build For a first computation we recommend that you choose a rather large number of PCs 1 2 or 1 3 of the total number of X variables for non spectral data around 10 for spectra so as not to miss the optimum in the variance curve Later on you may re calculate your model with fewer components e g one more than the optimal number so that the optimum still shows on the variance curve to limit the size of the model Model Size From the combo box choose between the three possible model sizes Full Compact and Minimum e The Full model which is the default contains all possible result matrices and thus can be opened in a Viewer and used in Prediction and or Classification on line or off line e The Compact model excludes the following result matrices from the model Eix X residuals Fiy residuals xWeighted and yWeighted weighted raw X and Y variables The Compact models have the same applicability as the Full models The Unscrambler Program Operation The Task Menu e 271 Camo Software AS The Unscrambler User Manual e The Minimum model only contains the result matrices needed for the Prediction and Classification analyses to work In addition t
413. t chapter contains a list of useful tips while the last one lists all menu options that allow you to transfer data or models between The Unscrambler and other applications Compatibility Overview of the systems The Unscrambler is compatible with Windows Compatibility The Unscrambler runs under Windows 95 Windows 98 Windows NT 3 51 or later Windows 2000 and Windows XP Data may be copied from The Unscrambler to other Windows based applications or the other way round by means of Copy Paste or drag n drop see section Useful Tips for Import Export for more details You may also copy graphs from The Unscrambler to other Windows based applications either directly or by saving a plot as bitmap which can then be opened in another application Overview of Compatible Formats The external file formats compatible for import and export to from The Unscrambler are listed in the table below Formats compatible for Import Export ky Format Import Import 3 D Export ASCH es Yes Data JCAMP DX es Yes ata Vision odel Tracker Yes Model INSAS Yes l Model IAPC es The Unscrambler Program Operation Compatibility e 39 GRAMS w D Matlab es Yes Data Guided Wave es Hitachi F3D Yes MVACDF Y es I Jata More details about Instrument compatibility are given in the Technical References chapter which is available as a PDF file from CAMO s web
414. t define the matrix you want to smooth in the Scope field The Define buttons in the Scope field launches the Set Editor where you can define new sample and or variable sets Then you enter the size of the segment to be used for smoothing i e how many adjacent columns should be used to compute the gaussian fitted value in the Parameters field Note The Gaussian Filter is supported in auto pretreatments Modify Transform Normalize Normalization is used to get all data in approximately the same scaling or to get a more even distribution of the variances and the average values It is a row oriented transformation that is to say the contents of a cell are likely to be influenced by its horizontal neighbors Chromatography data are usually normalized before analysis 222 e The Unscrambler Menu Options The Unscrambler Program Operation he Unscrambler User Manual Camo Software AS Dialog Normalize Normalize xj es Samples Cancel All Samples 100 v Define Defne SS Variables SC Spectra 100 7 Gene Define m Type C Area Normalization Unit Vector Normalization Mean normalization ik LK Abs Meant X i Maximum normalization ik X K 7 Mad Abel VI Range normalization LK esfLk 2 Maxi xti Mt zt Peak Normalization Var 21l ai Select In the Scope field in the Normalize dialog select the sample and variable sets
415. t shows the optimal number of components suggested by The Unscrambler You also have access to the variance as a numerical table and the warnings from the information field of the dialogs Wizards A wizard is a set of dialogs which are predefined by The Unscrambler as belonging together i e the choices made in one dialog naturally lead you to the next Dialogs in a wizard are connected with Next and Back buttons which allow you to switch back and forth between dialogs within the wizard The Next button in a specific dialog will only be enabled when you have filled in all necessary information with the help of the Back button you can however go back and change this information at a later stage In the last dialog of the wizard the Next button is substituted with the button Finish The Unscrambler Program Operation Dialogs and Wizards e 37 Getting Help in The Unscrambler When working in The Unscrambler you may get help on any feature by accessing the Help system or by using tooltips for toolbar buttons The Help System The Help system has been implemented to give you the help and advice you need when you are working with The Unscrambler Help is available on the following topics e Use of the dialogs e Use of the methods e Interpretation of plots Access the Help system at any time by pressing the lt F1 gt button or clicking on the Help button in the dialogs The Help file is automatically opened at the appropriate topi
416. t the default parameters will be re selected every time you make a selection from the Function drop down menu If your function is not listed but probably installed on the computer you can try to locate and register it by selecting Locate The Locate User Defined Transformations dialog will be shown The Unscrambler Program Operation The Modify Menu e 233 The Unscrambler User Manual Dialog Locate User Defined Transformations Locate User Defined Transformations x Below is a list of all registered COM classes on your computer that do not indicate that they belong to the User Defined Transformation category Acrok xch Document Register WARNING Be aware that trying to access register one of these may cause this program or the computer to yp hang or crash This depends on AUTOLICENSEMGR AutoLicenseMarObject 1 the component itself or the avifile consistency of the Registry AXSrvComp AXSrvComp database Ball2 Ball2 Daa Ball34x Be sure to save your other work Com mme applet BrowserAppletFrame or files before you attempt to COMCTI Imanel ist trl EI register one of these components hee If the UDT component is properly installed you will now find it on the list shown along with many other non UDT components installed Select the component in the list and click Register to let The Unscrambler try to register and access it as a UDT component A short information message will be displayed to tell you whether
417. t validation 312 Test Set Validation Setup dialog 265 Text Properties dialog 144 145 Text Warning List dialog 281 three way 313 three way data 303 tile horizontally 294 vertically 294 Tip of the Day dialog 300 toggle 3 D layouts 241 toolbar 30 173 tooltips 38 Tracker export model 79 import data files 61 training samples 312 transform gap segment derivatives 229 Norris gap derivatives 229 transform 219 baseline 231 center and scale 232 derivatives 229 MSC EMSC 225 multiplicative scatter correction 225 noise 228 normalize 222 reduce average 233 Savitzky Golay derivatives 230 smoothing 220 SNV 232 spectroscopic transformation 224 transpose 234 User defined 234 Transform Baseline dialog 231 transformations redo 243 undo 242 transpose 234 trend lines 159 regression line 159 target line 159 tri PLS The Unscrambler Program Operation Index e 327 analysis 273 prediction 284 Tutorials direct access from Help menu 299 U US data 53 71 U5 model 76 US results 53 UDA 312 UDI 65 register new UDI 70 unregister UDI 71 UDT 312 Components 234 235 Dialog 234 uncertainty limits 161 312 uncertainty test 160 313 Uncertainty Test dialog 265 undo 242 unfold 313 Unfold 3 D data 81 unimodality 313 unit vector normalization 222 223 unlock 84 unmark all 148 Unscrambler ASCII data export 71 Unscrambler ASCII files import 53 Upper Bound 89 See Lower Bound upper quartile 313 User Defined Import UDD 65
418. t variables The Unscrambler Program Operation The Modify Menu e 225 In the Multiplicative Scatter Correction dialog select the Sample and variable sets that define the matrix you want to correct in the Scope field If necessary press Define to enter the Set Editor and define new sets Then click the radio button Use existing MSC or EMSC Model if you want to be able to read an MSC or EMSC model from a file to transform your current data in exactly the same way as you did for an earlier data table This is useful if you want to treat different data tables in the same way e g new prediction samples Press the Find button to enter the Open File dialog if you do not remember the name of the MSC or EMSC model The default choice is to Compute and use new MSC or EMSC model by clicking this radio button You must then decide whether to make a full MSC model common offset additive effects model or common amplification multiplicative effects model in the Function field In addition to regular MSC you can also activate EMSC by clicking the check box Enable EMSC An extra tab becomes visible indicating the available options in EMSC Dialog Multiplicative Scatter Correction with EMSC options field Multiplicative Scatter Correction Scope Samples Selected Samples 100 z Variables All Variables 100 z C Use existing MSC or EMSC model ee Compute and use new MSC or EMSC model Msc EMSC
419. ter Useful Tips for Import Export for more information Supported Formats for Export The Unscrambler lets you export files as the following formats e Unscrambler ASCII e ASCII files e U5 data e JCAMP DX files e Matlab files e MVACDF files Export Unscrambler ASCII and U5 Data The dialog Export Unscrambler ASCII Export Unscrambler 5 lets you define the sets you want to export The Unscrambler Program Operation The File Menu e 71 Tick the box Export Y Matrix if you want to include Y variables in the file that you are exporting Then in the Sets field define the appropriate Sample and X Variable Sets If you have chosen to export Y matrix you must also select a Y Variable Set otherwise this option is not valid i e grayed out If you need to define new sets click the appropriate Define button to enter the Set Editor which allows you to define new sets or modify existing ones Export ASCII Files Many other programs can read ASCII files This export option therefore is very useful if you want to work with the data table in another program You enter the Export ASCII dialog where the Sets field lets you select the Sample and Variable Set you want to export Click Define to define new sets in the Set Editor Then specify whether to export as Flat or Flat Wide ASCII file in the File Format field Flat ASCII files are a string of numbers with no delimiters There are no characters to mark the different samples The
420. tes of the center samples are used to estimate the error Reference Samples Replicates of the reference samples are used to estimate the error Center and Replicates for both the center samples and the reference samples are used to estimate Reference Samples the experimental error COSCIND A forward significance testing method starting with the smallest effect and stopping when an effect is declared as significant The test statistics are related to the F statistic in the first choice but the p values are computed in a stepwise fashion Only the significance testing methods that are appropriate for your data are shown in the drop down list Effects Dialog Response Details Sheet The plot defined on this sheet shows the individual effects of the different variables and their interactions for one response at a time Choose between a plot of the individual effects as bars Line Plot or a Normal Probability Plot in the Plot Type field The Line Plot presents the main effects of the design variables and their interactions as bars or a curve F values or p values for checking significance of the effects can be added in separate windows The Normal Probability Plot shows the effects versus the expected effect values under the assumption that all effects are normally distributed around zero You can also choose to include a numerical table of the effects Dialog Effects Response Details sheet Effects OK Cancel War
421. th layout ON As a result e Former Secondary samples are now Variables e Former Variables are now Secondary samples e Variable names and Sample names are updated accordingly e Former Primary samples are still Primary samples The operation can be undone using Modify Undo Swap Sams amp Vars or CTRL Z The Unscrambler Program Operation The Modify Menu e 239 Modify Toggle 3 D Layouts Choose Modify Toggle 3 D Layouts or CTRL 3 for a quick change of the layout of a 3 D table This is a quick way to perform the 2x6 operations otherwise available under Modify Swap 3 D Layout and Modify Swap Samples amp Variables By using CTRL 3 repeatedly you will achieve the following results Result of toggling 3 D layouts repeated times on an OV or O V table Layout Layout times after CTRL 3 after CTRL 3 ov s pv sv 0 O V ps ss v 0 OV s sv pv O V ss ps v 1 ov sv s pv O V v ps ss 2 1 2 ov pv s sv ON v ss ps 4 OV pv sv s O V ss v ps OV sv pv s 5 O V ps v ss 5 eae eee O V s sv pv 7 OV ps v ss 7 O V pv sv s 8 ov v ps ss 8 O V pv s Sv 9 ov ss ps v 9 CV s pv sv 10 ov ss v ps 10 CV sv VE 11 ov v ss ps 11 Note Although all combinations are exhausted after 12 times you need to apply CTRL 3 16 times if you want
422. that describe the actions of toolbar buttons as you press them before releasing them while the right area indicates which user you are logged on as The size of the data table and the value of the current cell is displayed when an Editor is active It also tells whether you have read and or write access to the data in the active window The Unscrambler Program Operation The View Menu e 173 The Plot Menu Plot Line k The Line plot displays a data vector When you are plotting from the Editor you must mark the sample s or variable s you want to plot one sample variable gives a one dimensional plot specifying a range adds several line plots Dialog Line Plot Line Plot EN m Scope Variable set Cont Non Design Yars EI D Cancel il Help You enter the Line Plot dialog where you must specify a Variable Set if you have selected one or several samples or a Sample Set if you have selected one or several variables from the drop down list in the Scope field If necessary press the Define button to define new sets in the Set Editor Use Edit Options to change the layout of the plot Plot 2D Scatter The 2D Scatter plot shows two data vectors plotted against each other You enter the 2D 3D Scatter Plot dialog where you can specify the scope of the plot Note When you are plotting from the Editor you must mark the two samples or variables you want to plot before using the Plot comma
423. the corresponding sample numbers are displayed in the Initial Guess Samples field For more details on how to use the Define and Select buttons you may lookup the description of a similar sheet in PCA Dialog Samples Sheet p 259 MCR Dialog Variables Sheet In the Variable Set field select the variable set on which to perform the calculations Use Keep Out of Calculation if you wish to leave out some of the variables If you have enabled an initial guess based on some known concentrations the corresponding variable numbers are displayed in the Initial Guess Variables field For more details on how to use the Define and Select buttons you may lookup the description of a similar sheet in PCA Dialog Samples Sheet p 259 MCR Dialog Enable Initial Guess This field allows you to take into account any prior information you may have on the mixture constituents Initial Guess for Concentrations Use this option if some of the variables in the data table contain known concentrations of the mixture constituents or an approximation of those concentrations The Unscrambler Program Operation The Task Menu e 267 EN Camo Software The Unscrambler User Manual Type in the variable range directly or click Select to select relevant columns from an Editor window The corresponding variable numbers will be updated in the Initial Guess Variables field of the Variables sheet Initial Guess for Spectra Use this option if some of the sample
424. the experimental error Use at least two replicates if you know that your experimental results are likely to vary from time to time The Unscrambler only supports balanced designs so the number of replicates must be the same for all design points Note Replicates are not the same as repeated measurements Replicates mean that a new experiment is run using the same settings for the design variables as in a previous one while repeated measurements means measuring the response values for the same experiment several times The Center Samples are samples located in the center of the design This is only possible if all the variables are 2 level and continuous Center samples are used for curvature checking and for error variance estimation Use at least 2 preferably 3 or more center samples The location is defined by the levels of the factorial design variables The center samples correspond to the average mean of the different variables in the design In the field Reference Points you define samples which are incorporated for comparison A typical reference sample is a target sample a competitor s sample or a sample produced after to a given recipe The values of the design variables are not entered and are set as missing Design Details Next Dialog Once you are satisfied with your choice press Next to enter dialog Randomization Details General Design Wizard Randomization Details General Sometimes it is useful to perform some
425. thout decimals Note The Unscrambler uses the IEEE 754 standard data format Numbers are stored with six or seven significant digits If you select several decimal places and more than seven digits are shown in the Editor you should be aware that the last digits may be noise We recommend that you use the Autofit option to avoid this Use the Mixture component unit field to denote mixture components either as percentages 0 100 or as fractions 0 1 The default is to use percentages System Setup Dialog Viewer Sheet The appearance of the Viewer Stage and the plotting speed are affected by the properties in this sheet 92 e The Unscrambler Menu Options The Unscrambler Program Operation The Unscrambler User Manual Camo Software AS Dialog System Setup The Viewer Sheet System Setup l x General Editor Viewer Missing Data Printing r Clipboard format Tl Black text Bitmap T Black graphics C Picture Miscellaneous IV Disable bi info for large plots G e ei Ac plot when entering viewer IV Use numbers for blank names Define Colors Reset Options Cancel Apply Help The quality of the plot may be poor if you print a plot full of colors on a non color printer This is due to how the colors are mapped in the printer Tick the boxes Black text and or Black graphics in the Printing field if you want the text and or graphics to be plotted using black instead of gray tone
426. tical PC Previous Horizontal PC and Next Horizontal PC Note To access the Previous Vertical PC Next Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC buttons activate the Source toolbar View Toolbars View Source Previous Horizontal PC el The command View Source Previous Horizontal PC and its button Previous Horizontal PC is part of the PC navigation tool It allows you to switch to the previous PC along the horizontal axis of the active plot Note You may also use the lt arrow left gt key as keyboard shortcut for this command The command is available in the Viewer for PCA PCR MCR PLS1 PLS2 and three way PLS results on plots of the following types e 2D Scatter plots where the axes are Principal Components or PLS components e g Scores Loadings or similar e Line plot or 2D Scatter plot displayed for a specific number of PCs e g Influence plot Regression coefficients Predicted vs Measured or similar Note The command is not available for 3D Scatter plots Example Influence plot from a PLS regression model displayed for a model with four latent variables PCs The Unscrambler Program Operation The View Menu e 167 Q Camo Software AS The Unscrambler User Manual 6 Residual variance 05 115 4 4 080 00 0 tad 0 5 19 5 2 DEE aye adep 10 0 6 0 2 0 0 Tutor c pls2 PC 4 4 infiuence 10 0 10 0 Leverage
427. ting 203 204 re formatting redo 243 undo 242 registered licensee 6 13 21 301 registered users 302 regression 308 310 analysis 268 line 159 Regression Three Way PLS dialog 273 regression and prediction 199 Regression and Prediction dialog 199 regression coefficients 196 MLR 196 PCR and PLS 197 response surface 196 three way PLS 198 uncertainty limits 161 Regression Coefficients dialog 196 Regression dialog 268 269 regression methods MLR 268 270 PCR 268 269 PLS 268 269 PLS1 268 269 PLS2 268 269 regression overview 191 regression overview three way PLS 192 Regression Overview dialog 191 192 Regression Overview dialog for three way PLS results 192 remember plot settings 92 93 replace 127 replicate 310 replicates 116 117 reset viewpoint 157 residuals 186 general sheet 187 general sheet three way PLS 188 MCR fitting sheet 189 outliers sheet 188 patterns sheet 189 PCA fitting sheet 189 Residuals dialog 186 resolution 116 response surface 199 overview 202 response surface 277 response surface dialog general sheet 200 X variables sheet 200 Response Surface dialog 200 277 general sheet 200 X variables sheet 200 Response Surface Overview dialog 202 response variable 310 results all 289 Analysis of Effects 290 Classification 291 Data table 292 EMSC 291 general view 292 MCR 291 MSC 291 MSC EMSC 291 PCA 290 Prediction 291 Regression 290 Response Surface 290 Statistics 291 Results dia
428. tion The Plot Menu e 187 Choose between raw residuals or studentized residuals in the Y residuals field and between calibration or validation samples in the Samples field Finally select X variables Y variables or both in the Variables field The Y variable option is not available for PCA results Residuals Dialog General Sheet Three Way PLS Results In the Plot field on the General sheet you must choose which type of residual plot you want to display by selecting the appropriate radio button You have 7 choices as described hereafter Plots in the Residuals dialog General sheet Three Way PLS Results Plot type Description Y residuals vs Predicted Useful to detect outliers and or lack of model fit and systematic errors Normal Probability Y residuals A normal probability plot of the Y residuals which is useful to detect lack of model fit Y residuals vs Score A plot of Y residuals vs component scores which is another way to detect lack of fit in the model Influence Plot Squared residuals are plotted against leverage When a regression is performed you can use it both for X and Y variables The plot is useful for outlier detection Variance per Sample A plot of the average squared residual for each calibration sample taken over variables It is useful to detect outliers Variable Residuals A plot of residuals versus sample number for selected variables used to detect outliers and lack of model fit Sam
429. tives Norris Gap ssseessessssrsresessersre ceeeeeeesesaeeeeeaesaeeaeeaesees 228 Modify Transform Derivatives Gap Segment 229 Modify Transform Derivatives Savitzky Golay cceeeeeesseeesese ce eseeeeteeseeeeeaeeeeeees 229 viii e Contents The Unscrambler Program Operation Jnscrambler User Manual Camo Software AS Modify Transform Baseline cee eeeceesseesee ceeeeeeceeeeceeeeaeeeeseeseceeeeneesaseaeeeesaseaeeeaesees 230 Modify Transform SNV siaip renee e er E a rE E AEAEE EE EREE 231 Modify Transform Center and Secale eececeeseseeceeseseeeeeecaseseeeesaseesnerensseaeeeeeaseas 231 Modify Transform Reduce Average 232 Modify Transform Transpoe 0 cceccecesseeseeeesseeeeesseeeeeeesaeeeseeesaeeeseees caaseseasseeaesaeenseaes 233 Modify Transform User defined 0 0 eee ccceceeeeeceeceseseeseeeecaeseeeesaeeaesenessesseceesaeeneeaes 233 Modify Sort Samples sererai erae ann a n gevetagtyondegai netstat ehende nee 234 Modify Sort Samples by Setz 235 Select Sample Sets To Sort Select Variable Sets To Sort Dialog wo eee eee 235 Modify Sort Variables by Sets oo c cc esceseceseseescneeeecseeeceeecseecneesaesecesnessasssseesseeeeaesageas Modify Shift EE Modify Reverse Sample Order Modify Reverse Variable Order cctessciacesesscissenesesnsasesesbasassscavesustseczesisesssasanesdnecdapansoases Modify Swap 3 D Layout Modify Swap Samples amp Variables Modify Swa
430. to compute the smoothing function Modify Transform Smoothing Median Filter In the Median Filter Smoothing dialog you can replace each value in a row by the median of the values within a given segment centered on the point to be smoothed The Unscrambler Program Operation The Modify Menu e 221 The Unscrambler User Manual First select the sample and variable sets that define the matrix you want to smooth in the Scope field The Define buttons in the Scope field launches the Set Editor where you can define new sample and or variable sets Then you enter the size of the segment to be smoothed i e how many adjacent columns should be used to compute the median in the Parameters field Note The Median Filter is supported in auto pretreatments Modify Transform Smoothing Gaussian Filter In the Gaussian Filter Smoothing dialog you can replace each value in a row by a fitted value determined by a Gaussian filter function of its nearest neighbors In practice this amounts to averaging values within a segment of data points that have been wei ghted according to a Gaussian distribution function with os Dialog Gaussian Filter Smoothing Gaussian Filter Smoothing g x Samples T Cancel dl Samples 25 Define cma Variables Hep All Variables 262 DI Define m Parameters Segment size number of points 3 E Standard deviation is 2 First select the sample and variable sets tha
431. ton in a dialog you launch the Set Weights dialog 262 e The Unscrambler Menu Options The Unscrambler Program Operation nscrambler User Manual Camo Software AS Dialog Set Weights Set Weights x No Variable Name Weight W 1 Weight Let SE 2 Ink 1 SDev Cancel 3 Brightness 1 SDev 4 Scatter 1 SDev Help 5 Opacity 1 SDev 6 Roughness Passified 7 Permeability 1 SDev xf l Select all r Change Selected Variable A SDev B A 1 00 B 10 00 C10 2 Constant C Passify How To Use the Set Weights Dialog 1 First select the variables for which you want to set the weights in the top field You can either enter the number s of the variable s for which you want to change the weight s in the box below the list press Select to enter an Editor where you can find the variables you are looking for or press All to select all variables 2 Then you choose the appropriate weighting for the selected variables in the Change Selected Variable s field and press Update to update the selected weights in the list The following weighting options can be chosen from the Change Selected Variable s field at the bottom of the dialog e A Sdev B e 1 0 e Constant e Passify A Sdev B The weighting A SDev B with A 1 0 and B 0 0 i e 1 Sdev is called standardization and is used to give all the variables the same variance 1 If this is used all the variables are given the same chance to influ
432. tructure As a consequence source files in Matlab 4 or earlier format are no longer supported If you are using an older version of Matlab you may either upgrade to version 5 1 or use an ASCII format to exchange files with The Unscrambler Click OK to access the next dialog Import 3 D where you will specify your 3 D table layout see The Import 3 D Dialog for a detailed description File Import 3 D F3D Allows you to specify the parameters of the import from an F3D source file Hitachi The F3D Import dialog box displays a list of files from which you may import spectral data Each file is displayed with its File name and further characterized by the following parameters e Ex Start Ex End Ex Interval e Em Start Em End Em Interval e No Ex Var No Em Var If necessary click the Browse button close to the Look in field in order to access files from a different folder Once you have selected a file from the list information about the file contents is displayed in the File information field at the bottom helping you check that you have chosen the correct file Multiple selections are possible using Shift Click or Ctrl Click You may select one additional option e Add file name to sample name Click OK to access the next dialog Import 3 D where you will specify your 3 D table layout see The Import 3 D Dialog for a detailed description File UDI This menu option is only available when logged on as Supervisor see
433. ts are supported asd 001 002 3456 etc any number The Indico Import dialog box displays a list of files from which you may import Indico data File name number of X variables names of the First and Last X variables and Step are displayed for each file If necessary click the Browse button close to the Look in field in order to access files from a different folder Once you have selected a file from the list information about the file contents is displayed in the File information field at the bottom helping you check that you have chosen the correct file Multiple selections are possible using Shift Click or Ctrl Click You may select one additional option e Add file name to sample name Click OK to validate your choices and perform the import File Import BFF3 This option lets you import data from Brimrose instrument files Brimrose source files with the dat format are supported Dialog Brimrose Import Look in ambler 95012 DATA USE Test data Browse OK Cancel ES Help File name vars First var Last var Step Samples I Add file name to sample name File information File description lactose test 1 The Brimrose Import dialog box displays a list of files from which you may import Brimrose data If necessary click the Browse button close to the Look in field in order to access files from a different folder File name number of X variables wavelengths for the First and Last
434. ts button Previous Vertical PC is part of the PC navigation tool It allows you to switch to the previous PC along the vertical axis of the active plot Note You may also use the lt arrow down gt key as keyboard shortcut for this command The command is available in the Viewer for PCA PCR MCR PLS1 PLS2 and three way PLS results on 2D Scatter plots where the axes are Principal Components or PLS components Note The command is not available for Line plots or 3D Scatter plots Example Scores plot from a PCA model displayed for dimensions PC1 against PC3 The Unscrambler Program Operation The View Menu e 163 Camo Software AS The Unscrambler User Manual PC3 Scores Dein 3 2 1 0 1 2 3 4 Jams unscaled P X expl 58 11 After using the Previous Vertical PC button the scores plot is displayed for dimensions PC1 against PC2 PC2 Scores C4H3 i C4H2 em C3H3 3 2 1 0 1 2 3 4 Jams unscaled P X expl 58 28 See also Next Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC Note To access the Previous Vertical PC Next Vertical PC Back to Suggested PC Previous Horizontal PC and Next Horizontal PC buttons activate the Source toolbar View Toolbars View Source Next Vertical PC Ki The command View Source Next Vertical PC and its button Next Vertical PC is part of the PC navigation tool It allows you
435. ts has a fixed value The mixture sum can also be expressed as fractions with values varying from 0 to 1 306 e Glossary of Terms The Unscrambler Program Operation Mixture Variable Experimental factor for which the variations are controlled in a mixture design or D optimal mixture design Mixture variables are multi linearly linked by a special constraint called mixture constraint There must be at least three mixture variables to define a mixture design See Mixture Components Modes In a multi way array a mode is one of the structuring dimensions of the array A two way array standard n x p matrix has two modes rows and columns A three way array 3 D data table or some result matrices has three modes rows columns and planes or e g Samples Primary variables and Secondary variables Multi Linear Constraint This is a linear relationship between two variables or more A constraint has the general form A X A2 X2 An Xn Ao gt 0 or A X Az X2 An Xn Ao lt 0 where X are designed variables mixture or process and each constraint is specified by the set of constants Au An A multi linear constraint cannot involve both Mixture and Process variables Non Negativity In MCR the Non negativity constraint forces the values in a profile to be equal to or greater than zero Normal Probability Plot The normal probability plot or N plot is a 2 D plot which displays a serie
436. tton is used to erase all Log records for the current data file This button is disabled if the Supervisor has not authorized the option for the current user Enability disability of the Empty button is set in the Supervisor System Setup options In the Parameters column abbreviations are used to tell which transformations have been done See the Technical References chapter which is available as a PDF file from CAMO s web site www camo com TheUnscrambler Appendices for details about the abbreviations used in The Unscrambler Result File Only the transformations performed on the data table are displayed in the Transformations sheet These transformations are the same as recorded in the Log for data files Dialog Properties Transformations Sheet Tutor c bascorr signif Properties General Notes Transformations Model User Time Parameters DG 04 19 04 15 22 TYPE 0 FIRST You can select transformations for autopretreatment and or export to other model formats The currently selected transformations are marked with a red exclamation mark inside their icons 1 out of 1 transformation selected Select aen At the bottom of the sheet a Select button allows you to select the transformations that you wish to apply to new data as automatic pre treatments upon using the model at a prediction or classification stage Clicking the Select button opens the Select Transformations dialog see p 90 Properties Dial
437. tup and select the appropriate options in the Missing Data sheet You may select to be notified every time missing values are filled If so the following information dialog will appear every time this happens The Unscrambler Program Operation Working with Data Tables The Editor e 31 Camo Software AS The Unscrambler User Manual Automatic filling of missing values upon opening a data table in the Editor The Unscrambler E A Missing data was successfully filled in Note Unless you have chosen automatic filling of missing values the Information dialog below will appear every time you open a data table containing some missing values in the Editor You may check a box at the bottom of that dialog to turn it off in the future Dialog Information upon opening a data table with some missing values Information x sA The loaded data table contains missing values Not all OA calculations in The Unscrambler can handle missing data How do you want to proceed Ze Do Nothing C Open fill missing dialogue this time only C Configure automatic filling of missing data T Inthe future do not show this message Plotting from the Editor You can easily plot from the Editor Select the samples or variables you want to look at graphically and select the plot type you want from the Plot menu as seen in the figure below You can choose between several different plots depending on how many samples resp variables you have sele
438. two additional fields Primary Sam and Secondary Sam All three fields help you select the desired row in the most flexible way as shown in the figure below Dialog Go To for an O V 3 D table Edit Select Samples In the dialog Select Samples you can choose which samples you want to work with as the sample set Currently Selected Samples in your analysis You can extend or reduce the number of samples you work with at the moment Dialog Select Samples Selection method Select 7 OK ok Cancel Help Set fan Samples D I e Define C Levelled variable Variable Op Level fp JE fi o000 zj 132 e The Unscrambler Menu Options The Unscrambler Program Operation In the field Selection method a drop down menu gives you 3 options e Select Selects the samples you specify e Add Adds the samples you specify to the samples already selected e Remove De selects the samples you specify from the samples already selected In the next field Select Using you select 1 out of 3 available radio buttons Set Leveled variable Samples If you click Set the new samples can be chosen from a predefined set A set is defined by the Edit Set command in the Modify menu The button Define next to the drop down menu takes you to the dialog Set Editor where you may review all available sample and variable sets and choose between samples and variables Clicking Leveled Variable allows you to b
439. two variables or more ai Ki s X2 An Xn n gt 0 or ai Ki s X2 An Xn n lt 0 where X are relevant variables e g estimated concentrations and each constraint is specified by the set of constants dg Ay 2 Context Mixture Designs See Multi Linear Constraint Continuous Variable Quantitative variable measured on a continuous scale Examples of continuous variables are Amounts of ingredients in kg liters etc Recorded or controlled values of process parameters pressure temperature etc Corner Sample See vertex sample Correlation Loadings Loading plot marking the 50 and 100 explained variance limits Correlation Loadings are helpful in revealing variable correlations Cross Validation Validation method where some samples are kept out of the calibration and used for prediction This is repeated until all samples have been kept out once Validation residual variance can then be computed from the prediction residuals In segmented cross validation the samples are divided into subgroups or segments One segment at a time is kept out of the calibration There are as many calibration rounds as segments so that predictions can be made on all samples A final calibration is then performed with all samples In full cross validation only one sample at a time is kept out of the calibration Cube Sample Any sample which is a combination of high and low levels o
440. ual Dialog Design Wizard Enter Multi Linear Constraints Design Wizard Enter Multi Linear Constraints Lower limit multi linear constraints for process variables are of the form Al Basil 42 Oregano gt A0 Enter coefficients for each constraint New Delete Upper limit multi linear constraints for process variables are of the form A1 Basil 2 Oregano lt A0 Enter coefficients for each constraint A2 Ap 1 00 1 00 10 00 New Delete Information 1 Lower limit multiLinear constraint E 1 00 Basil 1 00 Oregano gt 4 00 1 Upper limit multiLinear constraint 1 00 Basil 1 00 Oregano lt 10 00 F cot lm _ The upper part of the dialog is used to define Lower limit constraints i e constraints of the form A1 DesignVariable1 A2 DesignVariable2 gt A0 The actual names of your design variables are displayed in text format The lower part of the dialog is used to define Upper limit constraints i e constraints of the form A1 DesignVariable1 A2 DesignVariable2 lt A0 The actual names of your design variables are displayed in text format Click on the New buttons to initiate the definition of new constraints Then type in the coefficients Al A2 A3 and the constant AO Use the Delete button to delete unwanted constraints At the bottom of the Information field you may view the defined constraints Click for example
441. up to and including the chosen number are displayed as a line plot in the lower sub view Plot Regression Coefficients You can plot the regression coefficients from a multiple linear regression MLR analysis or from a PCR or PLS regression analysis This is done in the Regression Coefficients dialog where you have to select the Plot Type line or matrix the Y variable and for PCR PLS and three way PLS the number of Components you want to include in your analysis Plot Regression Coefficients MLR and Response Surface Plotting the regression coefficients from a multiple linear regression analysis can be useful for interpreting the relationship between Y and all the X variables The regression coefficients are used to compute new Y values according to the equation for a linear model in addition the model may have interactions and or squares y bo LI b2X2 a bkXK f For MLR you can include the information listed below in the plot 196 e The Unscrambler Menu Options The Unscrambler Program Operation Information which can be included in an MLR regression coefficients plot Options Description The t values for the hypothesis that a regression coefficient is equal to 0 are used p values The p values corresponding to the observed t values are plotted Standard error The standard errors of the regression coefficients are plotted The dialog displays a preview screen Double click on this to
442. user defined analysis 286 user defined analysis UDA 313 User Defined Analysis dialog 286 user defined transformation 234 user defined transformation UDT 313 users register more 302 V validation 163 304 306 312 validation methods cross validation 260 leverage correction 260 test set 260 Validation Methods 260 270 cross validation 260 leverage correction 260 test set 260 validation samples 313 variable 313 active 308 category 31 continuous 31 131 Continuous 131 correlation 161 design 31 mixture 201 Mixture 129 131 200 Non design 131 Non Design 131 passified 308 process 201 Process 114 115 200 statistics 152 Variable Mixture 129 Variable Properties dialog 104 244 variable range error 262 Variable Statistics dialog 153 Variance dialog 282 variance plot 178 variances and RMSEP RMSE 180 X and Y variance 179 X or Y variance 178 X or Y variance three way PLS 179 X1 and X2 variance 179 Variances and RMSEP dialog 178 variances and RMSEP plots 178 vertex sample 313 view autoscale 154 back to suggested PC source 165 calibration source 163 correlation loadings 161 experiment sample sequence 154 explained variance source 170 graphical 154 Hotelling T2 ellipse 159 MCR message list 171 next horizontal PC source 168 next vertical PC source 164 numerical 154 outlier list 172 plot ID 34 158 plot statistics 158 point names 154 previous horizontal PC source 167 previous vertical PC sour
443. uto pretreatments See next note Note 3 The PCA and Regression dialog boxes include options for centering and scaling variables directly at the analysis stage It is recommended to perform centering and scaling at the model building stage in case you need the model for future prediction or classification The same centering and scaling options will be applied as when the model was built Note 4 Centering and or scaling the data more than once will not affect the structure of the data any further Consequently if you have applied a Center and Scale transformation to your data from the Editor the data may harmlessly be re centered and or re scaled at the modeling stage PCA or regression Modify Transform Reduce Average You can reduce the size of your data table by averaging samples or variables Averaging reduces uncertainty in the measurements and the effect of noise If you have the same number of replicates in your data table you can average the replicates to get one row for each sample Depending on whether you reduce along samples or variables the transformation is either column oriented or row oriented Choose whether to Reduce along Variables or Samples in this field in the Reduce Average dialog The number of adjacent samples or variables to be averaged must be given in the Reduction Factor field Note All defined sets will be adjusted according to the reduction you perform Reduce Average 3 D Data Reducing by averag
444. view table of the effects of all design variables on all the responses There are three sheets in the Effects dialog Overview Response Details and Effect Details Effects Dialog Overview Sheet On this sheet you can specify a table plot which shows an overview of significant main effects and two variable interactions on each response Each Y variable defines a column containing either NS or and signs The number of and signs in the plot indicates the degree of significance of the different effects as detailed below Effects significance limits P value Negative effect Positive effect gt 0 05 INS NS 0 01 0 05 The only choice you have to make on this sheet is which Significance Testing Method to use which is selected from a drop down menu The significant effects are found by comparing the effects with an estimate of the random variation This random variation can be found in five different ways as displayed below of which the first four choices are based on ordinary F testing with different error variance estimation Significance Testing Methods Method Besegten O Higher Order This method assumes that all effects except those specified in the model main effects Interactions HOIE main effects plus interactions are negligible They are used in the error term for experimental error variance estimates The Unscrambler Program Operation The Plot Menu e 203 Center Samples Replica
445. w Edit Insert Sample or Variable E al You can insert a variable or a sample in a given Editor A new column is inserted to the left of the column containing the currently active cell New rows are inserted above the active cell It is not possible to insert variables or samples between variables or samples in locked cells e g in a designed data table Inserted variables are always continuous An inserted sample is assigned the same type as the sample above it The cells are filled with m indicating missing data until you edit them Previously defined Sets are adjusted to accommodate the inserted range Insert Sample or Variable 3 D If your table has an Ov layout you may insert either Primary or Secondary variables as shown below Samples can be inserted as in 2 D data Dialog Insert Variable OV Insert Variable What would you like to insert A new primary variable resulting in 5 new columns Cancel C A new secondary variable resulting in 6 new columns Help The Unscrambler Program Operation The Edit Menu e 123 Camo Software AS The Unscrambler User Manual If your table has an OV layout you may insert either Primary or Secondary samples as shown below Variables or Category variables can be inserted as in 2 D data Dialog Insert Sample OV Insert Sample What would you like to insert C A new primary sample resulting in 5 new rows Cancel A new secondary
446. whether you want the text field to show the sample variable Numbers or the sample variables Names In the Insert field choose to insert specific samples specific variables or X general expression If you choose the Sample or the Variable options the drop down list is enabled and you may select the relevant object s from the list The available samples or variables are only those belonging to the Scope formerly selected in the Compute dialog Note The formula cannot contain mixed references to samples S variables V and X The Arithmetic Functions Trigonometric Functions Other Functions and Numbers fields offer buttons and are used in the same principle as for a calculator The More gt gt and Less lt lt buttons let you expand reduce the number of available functions Click Clear to refresh the expression Click Undo to undo the latest insertion in the expression text Clicking OK brings you back to the Compute dialog Modify Transform With this option you can transform your samples or variables to get data properties which are more suitable for analysis and easier to interpret Bilinear models e g PCA and PLS basically assume linear data The transformations should therefore result in a more symmetric distribution of the data and a more linear behavior if you have non linearities Note Transformations which may change the dimensions of your data table are disabled for 3 D data tables The Unscrambler Program Operation
447. xample 1 5 7 10 13 27 For more details you may lookup the description of a similar dialog in Statistics Dialog Samples Sheet Analysis of Effects Dialog X variables Sheet and Y variables Sheet In the Variable Set field you choose which Variable Set to use as X and Y respectively in the calculations For more details you may lookup the description of a similar dialog in Response Surface Dialog X variables Sheet and Y variables Sheet Task Classify The Classification dialog is used in the classification of new samples in comparison to existing known models The dialog contains two sheets Samples and X variables as well as some general options at the bottom 278 e The Unscrambler Menu Options The Unscrambler Program Operation Dialog Classification Classification x i X variables Sample Set fal Samples D I ze Define Keep Out of Calculation TTT i e Centered Models Nunberof PCs Model Method Max Sugg Use Pret PLS1 Result PLS1 2 0 1 Cancel dr Help Variance Pretreat ii Number of PCs to Use l Add Model Remove Model Classification Dialog Samples Sheet Choose which samples to use in your calculation in the Sample Set field and enter any samples which you do not want to include in the analyses in the Keep Out of Calculation field For more details you may lookup the description of a similar dialog in Statistics Dialog Samples Sheet p 253 Class
448. xes of a 2D scatter plot etc Together with the Plot ID this gives you a good overview of the data contents of the current sub view Identification window Plot Identification Cava Joe L ns Less Row Les Pore 1 Abscissa PCA Resut Tai PDC OI 1 36 Ordinate PCA Result Tai PC_02 1 36 By default the identification window is docked to the bottom of the Unscrambler workspace You may drag it with the mouse and move it around Drop it somewhere inside the workspace and it will float over the Viewer as a standalone window Window Warning List This toggle option available in a result Viewer window shows or hides a list of all the warnings that The Unscrambler generated during computation of the analysis The list is displayed in a dockable window These warnings are given in plain English and derive from the thousands of tests that were performed during the model computations The Unscrambler Program Operation The Window Menu e 295 The Unscrambler with the Warning List displayed as docked window at the bottom W The Unscrambler Const PLS2 E File Edit Yiew Plot Task Results Window Help SCE TE SE Ta Slope Offset Cor 0 937836 0 243107 0 968419 0 925159 0 266242 0 958907 Press Outliers to look at a list of the tests that were positive The outlier list gives a detailed view of the outlier and leverage detection tests By default the warning list window is docked to the bottom of the Unscrambler workspace You m
449. y of File New Design Modify existing design after giving it a name with File Save 118 e The Unscrambler Menu Options The Unscrambler Program Operation The Edit Menu Edit Cut Al or Ctrl X This option removes the selected range either data in the Editor or a plot in the Viewer and places it on the clipboard Anything that you place on the clipboard remains there until you replace it with a new item You can restore your selection from the clipboard using the Paste command Edit Copy or Ctrl C With this option you copy the selected range to the clipboard overwriting its previous contents The selected range is with other words not removed from its original place Use the Paste command to copy your selection to a new location Edit Clear This option is only enabled from a Viewer and clears it of all its contents thus allowing you to make a new plot in the same Viewer Edit Paste El or Ctrl V Lets you insert a copy of the clipboard contents at the insertion point The command is not available if the clipboard is empty or the selected range cannot be replaced When pasting data from another Windows application the Select Paste Method dialog pops up Dialog Select Paste Method Select Paste Method Insert as 3 new columns OK C Insert as 12 new rows Cancel Overwrite existing data Make your choice and click OK The Unscrambler Program Operation The Edit Menu e 119 Camo Software AS
450. yed 300 e The Unscrambler Menu Options The Unscrambler Program Operation he Unscrambler User Manual Camo Software AS bd D Dialog Tip of the Day Tip of the Day V Did you know To reduce the file size of your PCA PLS or PCR model select model size Compact or Minimum in the analysis dialog Below the text window a tick box Show tips on startup allows you to enable disable the Tip of the Day function on startup Use the Next Tip button if you want to browse through the tips database Close to close the dialog box Help About Tells you which version of The Unscrambler you are using and copyright notices Press Licensee to see who the licensee is and to get information regarding the number of registered and simultaneous users Help Change License The Change License dialog allows you to modifying your current license of The Unscrambler by typing in a new activation key Use this feature for instance to upgrade from a trial installation to a full version of The Unscrambler Dialog Change License The Unscrambler Change License x Ubtain Activate Unscrambler machine code Unscrambler activation key Modifies current License Cancel Note This command is only available when logged in as Supervisor To log in as Supervisor follow the instructions in the System Configuration chapter p 22 The Unscrambler Program Operation The Help Menu e 301 Camo Software A
451. yed in sub views 2 and 3 The number of components must be selected for each sub window e Inthe Components field specify along which components you want the results to be plotted e Inthe X variable Mode field choose between Primary and Secondary X variables e Finally make your choice between plotting X Y or X and Y variables in the Variables field Loadings Loading Weights Dialog 4 x 2D Scatter Sheet This sheet lets you specify four 2D Scatter plots The number of components must be selected for each sub window Specify which components you want to see in the Components field Finally make your choice between plotting X Y or X and Y variables in the Variables field Specific to Loading Weights For Y variables Loadings will be plotted since Loading Weights exist only for X variables Choosing X and Y variables will produce a plot of X Loading Weights and Y Loadings Loadings Loading Weights Dialog 4 x 2D Scatter Sheet Three Way PLS Results This sheet lets you specify four 2D Scatter plots which are displayed in sub views 4 to 7 The number of components must be selected for each sub window e Inthe Components field specify along which components you want the results to be plotted e Inthe X variable Mode field choose between Primary and Secondary X variables e Finally make your choice between plotting X Y or X and Y variables in the Variables field Plot Residuals Residuals are the dev
452. you have to choose between single double or quadruple plots giving 1 2 or 4 components in the Components field In each case you also need to select the correct number of components The Unscrambler Program Operation The Plot Menu e 193 Dialog X Y Relation Outliers X Y Relation Outliers OK Cancel Help Plot Y Relation Outliers Components Suggested no of components is 4 e Double Quadruple VE zs 2 I EN I Change the size of the plot by double clicking in the preview screen in the dialog Plot Predicted vs Measured In this plot the measured Y values and the predicted Y values are plotted This is useful to check the quality of the regression model fitted to the data Predicted values should be as similar as possible to the measured values Dialog Predicted vs Measured Predicted vs Measured E Cancel ble Help CAMO r Plot Type CS Predicted vs Measured C Predicted and Measured I Include Table Sample IV Calibratior 1 a D I Validation Y variable In the Predicted vs Measured dialog choose which plot you want to make in the Plot Type field You have two options e Predicted vs Measured which plots the predicted Y values versus the measured Y values for all samples 194 e The Unscrambler Menu Options The Unscrambler Program Operation e Predicted and Measured which is a plot of Y values with one curve showing the measured values and
453. you want to display in the plot You must select both X and Y variables To select All variables click this button and to enter an Editor where you can select variables click Select Tick the Total box to display the total variance at the same time You can decide whether you want the variance to be plotted for calibration or validation samples respectively or both by ticking the appropriate box es in the Samples field Calibration variance measures the fit of the model to the calibration data The validation variance is an estimate of the fit of the model to new data not present in the calibration and is based on the validation Plot both measures to obtain full information about model fit You can later change that choice by using View Source Variances and RMSEP Dialog X1 and X2 variance Sheet This dialog sheet enables you to plot Primary X X1 and Secondary X X2 variances simultaneously as a double plot Variance is plotted against component number You can choose between calibration and validation samples or both Explained variance is selected by default you can later change that by using View Source The Unscrambler Program Operation The Plot Menu e 179 Dialog Variances and RMSEP X1 and X2 variance Sheet Variances and RMSEP Plot Variance Plot 1 Samples Prix 1 10 All IV Total IV Calibration I Validation i Secx 1 5 All IV Total Prix fro 4 ass e In the fiel
454. your matrices contain the same number of variables row wise import or samples column wise import If the matrices contain sample and or variable names you can import them as well by sele cting the appropriate option box Note Some matrices have more planes one for each PC calculated Normally you should choose the same plane for all The planes may be PCs groups etc according to which matrix you want to import Several planes can be imported at once by entering a range e g 3 6 If you import several matrices and want to import another plane than the first you specify the correct plane number after you have marked the first matrix File Import Unscrambler Data You can read data files made by earlier versions of The Unscrambler into the Editor to study their numerical values by selecting the file type and file you want to import in the Import dialog You then enter the Import from X dialog where X represents the data file type You select the matrices you want to import from the matrix list See the Technical References chapter which is available as a PDF file 52 e The Unscrambler Menu Options The Unscrambler Program Operation from CAMO s web site www camo com TheUnscrambler Appendices for a list of all matrices used in The Unscrambler First you choose whether you want to import rows or columns by clicking the appropriate radio button Then select the matrices you want to import from the fil
455. zerpog Desan E Jnclude Interior Points Axial Design T Include End Points cmo zs The Reset button brings back the system defaults for all fields Note In a Simplex Lattice Design of degree gt 2 all interaction and square effects can be included in the model For degrees 1 and 2 only interactions between process variables and mixture variables are allowed Design Type Mixture Next Dialog After choosing a mixture design type in the Design Type Mixture dialog you enter the next dialog Design Details Design Wizard Generate Design This applies to D optimal designs only This dialog gives you access to D Optimal design generation Note that it contains a radio button Full Factorial Combination which is related to the case of a Mixture Design this button is disabled in the case of a Non mixture design The Unscrambler Program Operation The Design Wizard e 113 e AS 1e Unscrambler User Manual Generate Design D optimal Non mixture Design Wizard Generate Design D Optimal non mixture situation Design Wizard Generate Design xi Choose if you want the mixture part of the design to be combined with all levels of the process variables or if you prefer to reduce the number of experiments with a D optimal algorithm Full Factorial Combination Design Paints 1 Replicates D Optima fis 4 Design Points x 1 Replicates 3 Number of Center Points Above you can choose the number of points for your

The Unscrambler Program Operation

Contents

Download Pdf Manuals

Related Search

Related Contents