USER MANUAL GALLITO 2.0 64bits
Contents
1. me canes x E ETRE SSS LS c azkargorta club L gt M a callo Ti pne Peete p gt i LL o Se Em M clasificar epi magi goleador rojiblanco Avot clifford que bakero rugby2. supuesto incorrectas discuten se as acuerdo coe 3 parece ser hay intereses politica cientifica sostienen teorias diagonal matrix estado decir abogado si intereses politica cientifica ceden NOCHO PC tuberia santrock aes correcta dejar a existir buena parte pr spero campo osis Space features 2222222222 3 222222222 OCHO PC tuberia santrock dicho cesando abogado excusa decir entendimiento hipnosis OCHO PC tuberia santrock t rminos teatro apuntado tiene promisorios campos investigaci n l neas hipnosis perspectiva pone terapia relaci n contextos convienen lado estar a afinidad ret rica Modo sofistica investigaci n conducta verbal pasando Binary sugestionabilidad estar a aire familia terapia guarda teatro hipnosis tan pone relieve u Load 41572 2 2 2 C Serialized colegio oficial psic logos espa a ver texto pagina sucesivas datos completos profesional realizaron pruebas informe aparece nombre apellidos psic logo pone contacto instructor indicarle era intenci n ten a formaci n psicologia jur dica suficiente trataba compromiso indirecto tenia familiar se ora c dej guiar indicaciones abogado sta instructor indica argumentos formalice poder ser evaluados comisi n deontol gica efectivamente dentro plazo remite pliego alegaciones folios perito es pues es autocalifica repite innumerables veces hab a obrado buena voluntad elaborar informes hab a tenido cuenta intereses ni a
3. MATRIX E Occurence matrix to t rminos teatro apuntado tiene promisorios campos investigaci n l neas hipnosis perspectiva pone terapia relaci n contextos convienen lado estar a afinidad ret rica sof stica investigaci n conducta verbal pasando sugestionabilidad estar a aire familia terapia guarda teatro REDUCED MATRIX READY TO QUERY colegio oficial psic logos esparia ver texto p gina sucesivas datos completos profesional realizaron pruebas informe aparece nombre apellidos psic logo pone contacto instructor indicarle era intenci n ten a formaci n psicolog a jur dica suficiente trataba compromiso indirecto tenia familiar se ora c dej guiar indicaciones abogado sta instructor indica argumentos formalice poder ser evaluados comisi n abren destacar folios responde preguntas instructor propone abreviada conteste expuesto comisi n deontol gica delegaci n colegio 22 24 S oa OM Ft BO 4 2 Operations on semantic spaces Once a semantic space is created and loaded operations can be performed on it These operations can be comparing two terms comparing two documents identified by a number comparing two free texts entered by the user extracting semantic neighbors with simple or corrected cosines or with simple or corrected predication The space can also be saved in a hard disc directory to be loaded at another time Space properties In this option the properties of the
4. reitera vez mala Load classes list 22 30 ES a OQ EID 05 07 2012 Info Project loaded Do you want to go to the main panel Select Accept and the space will be loaded in the memory to perform the operations described above 4 3 Batch processing Actions can be performed by batches specifying in a file the terms and characteristics of the operations to be performed 4 3 1 Neighbor batches The first n semantic neighbors will be extracted from a number of terms In the expandable menu the following screen will appear where the number of neighbors the directory where the file specifying the terms and the file itself are specified The file will have the following format match football note The same process will generate a file per term specifying the neighbor cosine and vector length 2 Neighbor patches Neighbors 3 Directory CAUsers guille Documents E Term list in the same directory Generate 4 3 2 Similarity matrices A matrix will be extracted to compare the neighbors with each other This square matrix has ones in its diagonal and each cell represents the cosine of two neighbors The directory for the reference file and the file name itself will be specified Similarity matrices will be generated in that same directory The file format is the following match 200 football 300 note 300 Where a 200x200 will be extracted in the first c
5. spaces on which the user is working can be viewed Some of the indices will be disabled in the applications aimed at large linguistic corpora Space gt Properties Dimensions Term similarity average Term Similarity SD Document similarity average Document similarity SD Order 1 Index Order 2 Index Preprocess Comparing two terms Two specific terms can be compared by means of their cosine or the Euclidean distance between them Queries gt Term Term V Term Term T2 nerviosismo Results vector length 1 133539 Similarity 0 3990712 Comparing two documents identified by a number Two specific documents can be compared by means of the cosine or the Euclidean distance between them Queries Document Document describir programas constructivistas sociales est soluci n blemas gran parte instrucci n escritura Similarty 0 205256185405275 Comparing two freely produced texts Two free texts can be compared by means of their cosine or the Euclidean distance between them In large size spaces this process will take a few seconds Process progress will be shown by means of a status bar Queries gt Free texts respecto aclaraci n mitos documentada vigorosa exposici n capafons mayor honestidad intelectual profesional mira deja hipnosis pr cticamente placebos trucos sentido dejar a apreciar cierta convergencia desmitificaci n capafons desenmascaramiento ensayado TR UR dera lesm
6. used in the clases txt must be provided Load classes list Once the parameters for corpus creation have been selected open the Process tab and click on Start Then the processes will be highlighted as they are executed The total process can be slow and may even take days depending on the options and the text corpus Process Rem Add Corpus Matrix Save tab Load tab CLEANING TEXT ADDING OR REMOVING COMPILING TERMS GENERAL MATRIX Decurence matrix to REDUCED MATRIX READY TO QUERY 5 ST The final process ends with the following message Proyecto1 Aviso Matriz factorizada calculada Listo para consultas Aceptar Once this warning is accepted terms and documents will be loaded on the right hand side and it will be possible to perform operations on the semantic space Main menu ESO File Spaces Structures Queries Special queries Export Outputs batches License Process Rem Add Corpus Matrix Save tab Load tab razones requeria teoria esclarecedora hipnosis estado difuso profuso confuso teor a propuesta aun mejorada parece vaya cr dito fama influencia fascinaci n es mejor si simplemente ignorara careciere inter s iba hipnotizar si hubiera o do hablar hipnosis lugar profusi n cl nica cumple funci n renovar entusiasmo cl nicos aplican t cnicas est n sujetas necesitan CLEANING TEXT 23 ADDING OR REMOVING COMPILING TERMS rz GENERAL
7. Batch processes similarities graphs neighbors e Matrix output in plain text e Pajek format output for term visualization e Essay evaluation e Text Cohesion sentence sentence paragraph paragraph etc e Dimensions Interpretation 2 INSTALLATION 2 1 Requirements In order for Gallito 2 0 to work properly the following components must be installed they are included in the installation CD and can be downloaded from the Microsoft website e Windows 64 bits operating system Windows 7 or Windows Server e Microsoft SDK 4 included in the installation CD e Microsoft Visual C 2010 Redistributable Package 64bits e Writing permissions to write in the installation directory Some knowledge of Latent Semantic Analysis LSA and its applications is also required 2 2 Installation procedure Access the installation folder and execute the Setup exe program e The following screens will appear to guide you through the installation process 2 2 1 Step one Welcome screen to product installation specifying the product to be installed Gallito 2 0 Click NEXT to continue installing or CANCEL to exit the installation process The installer will guide you through the steps required to install 2 0 64Bits on your computer WARNING This computer program is protected by copyright law and international treaties Unauthorized duplication or distribution of this program or any portion of i may result in severe civil a
8. File Spaces Structures Queries Special queries Export Outputs batches License razones requeria teoria esclarecedora hipnosis estado difuso profuso confuso teor a propuesta aun mejorada parece vaya Process Rem Add Corpus Matrix Savetab Loadtab f Load project abogado primer lugar difusi n folkl rica hipnosis plaga errores Tenk mitos cumple funci n dare prestigio quiz renombre erm list cr dito fama influencia fascinaci n es mejor si simplemente NOCHO PC tuberia santrock ignorara careciere inter s iba hipnotizar si hubiera oido hablar hipnosis lugar profusi n clinica cumple funci n renovar Doc List entusiasmo cl nicos aplican t cnicas est n sujetas necesitan innovaci n NOCHO PCXuberiasantrock 3258 respecto aclaraci n mitos documentada vigorosa exposici n Doc matrix capafons mayor honestidad intelectual profesional mira deja WOCHO PCtuberia santrock e hipnosis practicamente placebos trucos sentido dejaria apreciar cierta convergencia desmitificaci n capafons Term Matrix desenmascaramiento ensayado casos hipnosis pierde NOCHO PC tuberia santrock solvencia mientras ificaci n resuelve ayuda adjunta NOCHO PC tuberia santrock t cnicas desenmascaramiento disuelve s figuras ret rica global weight teatro tercer lugar confusi n cient fica cumple funci n dar trabajo investigadores sector publican teor as discutibles NOCHO PC tuberia santrock
9. OTR Aa Version 2 Date 04 06 2012 1 Oficina de Contact sjansa pas uned es ransjerencia T Pa ce Name Document Type User_manual_v2 doc User Manual USER MANUAL GALLITO 2 0 64bits CONTENTS 1 2 INTRODUCTION INSTALLATION SUMMARY OF POSSIBLE ACTIONS APPLICATIONS 1 INTRODUCTION The aim of this document is to explain how Gallito 2 0 works Gallito is a tool suitable for two domains Research Makes it possible to process language samples and extract samples for psycholinguistics experiments such as term entropy frequency measured through vector length similarity between terms lists of semantic neighbors All of this is very useful to exercise experimental control or even to do research into the effect of this kind of variables on processing Technology For text categorizers in any domain In fact Gallito 2 0 is currently being used as the first module in cloud applications that categorize customer service calls It can also be used to visualize useful information for marketing departments The functionality associated with this application is Creation of semantic vectorial spaces from texts e Calculation of term significance functions Entropy or IDF e Calculation of vector norm for every term e Lists of semantic neighbors of a term e Similarity between terms e Similarity between existing documents e Similarity between documents that do not exist in space pseudodocuments e
10. ally 1 should be established This information is entered in the Corpus tab in the central panel Documents Documents character separation sentence separation character separation E sentence separation A document is a minimum of 1 sentences A document ls a minimum of 1 sentences A document minimun are El words 16 words A document minimun are Rem words that do not appear in at leasfg document Rem words that do not appear in at least ZZ document A document is a minimum of words This is the minimum number of terms for a document to be analyzed This information is entered in the Corpus tab in the central panel Remove words that do not appear in at least documents The minimum number of documents in which a specific term must appear to be included in the analysis This information is entered in the Corpus tab in the central panel Hem words that do not appear in at leas 6 a document Remove Literal occurrences of each of the structures presented are removed This information is entered in the Remove tab in the central panel e Generating a stop list 1 In the Structures Select drop lists select and enter the structures that you wish to remove In the left hand side you will find the structures composed by more than one term In the right hand side you will find simple or single term structures Structures composed by more Simple or single term
11. ase 300x300 in the second one etc E Similarity matrices IES n EM CXLlsers quille E File 4 3 3 Similarity between pairs This will generate the similarities between a number of term pairs The directory for the reference file and name of the file itself will be specified The file format will be the following tremendous action dog cat share stock Term list in the same directory lst bd 4 4 Standard output 4 4 1 Matrices to plain text files The Export drop menu includes the option Matrices to txt Just enter the directory in which you wish to generate the files which will be the US SV and S matrices as well as the weights assigned by the calculations to each word log entropy or log idf and the vector length for each term 4 4 1 Term to Pajek file For visualization purposes there is a Term to Pajek option in the Export drop menu From this process a file is extracted that will serve as input for the Pajek program to generate visualization networks Directory C Users guille Generate Pajek is available at http vlado fmf uni lj si pub networks pajek and can be downloaded for free This output will use cosines as similarity and vector length as node size By entering the input graphs such as the one below can be obtained nataci n Ibutraquedto liaquotot ey centrocampista 9 seleccionador lompi A ON ca m limi toria
12. fic terms can be extracted by various methods Queries gt Representation Bl Tor Representation ca p trastomo 12 3545 732116659 depresi n9 85341 167449951 suefio 64798831939697 sintomas9 54744815826416 trastomos9 42812252044678 psicologia9 20697784423828 ef 15685081481934 ansiedad9 1335391998291 hipnosis9 03021 144866943 criterio 58086242675781 salud8 5591402053833 pacieriteB 49645137786865 Saving a semantic space A semantic space can be saved in the hard disc and reloaded when necessary with no need to creating it against Save tab Then a name and path to save 7 variables must be selected We recommend that you save all spaces in the same directory which can be created in the same directory browsers as each of the variables At the end select the format we recommend binary and click on Save 9 Main menu o File Spaces Structures Queries Special queries Export Outputs batches License Process Rem Add Corpus Matix Save tab Load tab a razones requeria teoria esclarecedora hipnosis estado difuso profuso confuso teoria propuesta aun mejorada parece vaya tener xito siguientes razones argumentadas haciendo Save project Term list El E peste innovaci n Doc list Term matrix t cnicas desenmascaramiento disuelve s figuras ret rica E teatro tercer lugar confusi n cient fica cumple funci n dar trabajo inve
13. ighbors similarity by pairs similarity matrices To do this the application has a single form or control panel with tabs and drop lists The application has two basic functionalities e Generation of a semantic space e Operations on a semantic space Calculate Similarities Save Space Load Space etc 4 APPLICATIONS When the application is launched a small presentation is deployed followed by the control screen From then on you can create a new semantic space or load an existing space that has been saved in the hard disc The screen that appears is the following Main menu File Spaces Structures Queries Special queries Export Outputs batches License Process Rem Add Corpus Matrix Save tab Load tab 222 1 1 INS X M 21 Registered OTRI UNED Info www elsemantico com Oro fo ivi You can start to perform operations using the tabs and drop lists 4 1 Creating a semantic space To create a semantic space a corpus Is needed This corpus will be called reference corpus and will be a text plain file txt extension The contextual window can be separated by a character or simply by natural language sentences In the former case the separating character usually have to be between two documents Los archivos planos constituyen la forma m s b sica de una base de datos Los archivos planos incluyen un campo por cada uno de los elementos que se desea
14. iificaci I se as intereses pol tica cient fica ceden pol tica correcta dejar a existir buena parte pr spero campo hipnosis dicho cesando abogado excusa decir entendimiento hipnosis t rminos teatro apuntado tiene promisorios campos investigaci n l neas hipnosis perspectiva pone terapia relaci n contextos convienen lado estar a afinidad ret rica sofistica investigaci n conducta verbal pasando sugestionabilidad estar a aire familia terapia guarda teatro hipnosis tan pone relieve Resultados vector length 1 170221 Similarity 0 503112 Extraction of semantic neighbors The semantic neighbors of a specific term are extracted by various methods cosines corrected cosines predication corrected predication Neighbor trees will be displayed as term neighbors are extracted In addition the number of neighbors to be extracted can be selected The results can also be exported to Microsoft Excel Queries gt Semantic neighbors 9 Semantic neighbors Sa a sate Term trees Term trees Term nerviosismo Term fobia a la sangre Measures Measures Cosines C Cosines eee C Corrected cosines Predication Predication Corrected predication Corrected predication 5357 nerviosismo 12771 Extraction of the most representative terms in the semantic space The semantic neighbors of speci
15. inguistic adjustment Log Entropy t Log IDF t No aplicar ajuste e Normalizing the U Matrix The U matrix of terms extracted in the SVDF process is normalized before weighting it applying the weight of each dimension In this way the effect of term frequency is countered This information is entered in the matrix tab in the central panel Normalization Normalizing the U Matrix Reference corpus the path of the text file where the linguistic corpus is stored in a legitimate format Click on the button to browse through the directories This information is entered in the Corpus tab in the central panel Heference corpus n The chosen corpus can be selected in the standard Windows browser DIRECTORIO OCHO LSA gt corpus CIE Buscar CI Organizar v Nueva carpeta Favoritos 4 WE Escritorio J Bibliotecas Nombre CIE10 bd 19 CIE 10 bt bak 11 Clasificaci n de trastornos mentales CIE Fecha de modifica 05 07 2012 2 49 28 06 2006 19 54 28 06 2006 19 53 20 Documentos Im genes Musica im V deos 8 Grupo en el hogar i guille de gconf de gconfd gnome2 de gnome2 private Nombre Document separation The documents can be separated by a character or simply by natural language sentences In the former case by the separating character usually In the second case the number of sentences constituting a document usu
16. n contemplar La redundancia de elementos es una caracteristica de estos archivos La base de datos relacional soluciona la redundancia en los datos Son frutos largos y con sabor La recogida sera buena si ha tenido una buena base como semillero Los frutos son de color verde En la recogida es parecida a los demas frutos largos In the second case a well established sentence dot separation is mandatory but a special format is not needed To create a semantic space the following parameters are required e No of dimensions or Accumulated singular value The number of dimensions must not surpass the total number of documents The accumulated singular value is expressed as a percentage with no sign This percentage reflects the saved dimensionality that is to say the dimensionality percentage that will be maintained In this way 40 refers to the number of dimensions corresponding to said dimensionality In extremely large corpora it will not be possible to calculate this percentage and so 300 dimensions will be used These data are entered in the matrix tab in the central panel Matrix features f Dimensions 2 C Accumulated singular value Linguistic adjustment This option refers to the significance calculation for every corpus term Log Entropy or log IDF can be selected as well as the absence of these calculations This information is entered in the matrix tab in the central panel L
17. r criminal penalties and will be prosecuted to the maximum extent possible under the law Cancel 2 2 2 Step two In this screen you must select the directory to for product installation which by default will be C Program Filestelsemantico comiGallito_2 0 64Bits The installer will install 2 0 to the following folder To install in this folder click Nest To install to a different folder enter it below or click Browse Folder C Program Filezelzemantica com Gallita 2 0 b4Bits Install Gallito 2 0 amp 4Bits for yourself or for anyone who uses this computer Everyone Just me If you want to install the product in a directory other than the default directory select it by clicking BROWSE in the installation window Clicking this button you will access to a standard Windows directory browser through which you will be able to select the directory where you want to install Gallito 2 0 see following image Browse _1 2 0 B4Bits 2 palabrasClases Quitar Folder Filez elsemantica com allita 2 0 Once you select the installation directory click NEXT to continue the installation process 2 2 3 Step three In this step the user is warned that clicking NEXT will launch product installation If you click BACK you will be able to re enter the installation information 7H Gallito 2 0 64Bit The ins
18. stigadores sector publican teor as discutibles Global weight supuesto incorrectas discuten corrigen se as acuerdo coe Mo parece ser hay intereses pol tica cient fica sostienen teor as ES estado decir abogado si intereses pol tica cient fica ceden pol tica correcta dejar a existir buena parte pr spero campo Diagonal matrix ipnosis JE poe sugestionabilidad estar a aire familia terapia guarda teatro hipnosis tan pone relieve 15 Modo colegio oficial psic logos esparia ver texto p gina sucesivas datos completos profesional realizaron pruebas informe Binary Save nombre apellidos psic logo pone contacto instructor C Serialized indicarle era intenci n ten a formaci n psicolog a jur dica suficiente trataba compromiso indirecto ten a familiar se ora c dej guiar indicaciones abogado sta instructor indica Save project automaticatly 2 2 cel 05 07 2012 QS na Loading a semantic space In the same way a semantic space can be loaded from the hard disc with no need to create it again Load Tab A name and route to locate the 7 variables of the space to be loaded must be selected In the same directory browser as each of the variables you can browse and select the path for the relevant variable At the end select the format in which the space was saved we recommend binary Mainmenu X AAA A M X l
19. structures than one term 2 In the Remove tab check the Additional box pronouns Verbs Iw Additional e Creating a space with only one go list The procedure is similar to the previous one In the Remove tab the Additional box is checked and in Structures gt Select the structures in the go list are selected The Add exclusively option must also be checked Remove t Add exclusively Using this method you can select the remaining structures so that they are included in the analysis for instance an analysis with function words and adverbs and the selected additional structures This information is entered in the Remove tab in the central panel e Replace terms by classes To reduce the variability of some words it can sometimes be interesting to include terms in categories and use those categories as terms For instance all mobile telephone brands could be included in the ClassMobile category In this way in the process both Nokia and Alcatel would be treated as the ClassMobile term whose occurrences would increase This information is entered in the Remove tab in the central panel Classes i Replace terms by classes Class definition would be included in clases txt in the directory C Program Files elsemantico com Gallito_2 0 64Bits palabrasClases If you want to load that has been previously processed using class definitions it must be specified in the Load tab The definitions
20. taller is ready to install Gallto_2 0 b4Bitz on your computer Click Nest to start the installation 2 24 Step four The window shows installation progress as a percentage and the files being copied to the machine on which the product is being installed Once the installation process is 10096 complete you will access the next step see next section after being informed that the necessary information for the application to work correctly has been entered Clicking CANCEL in the progress window halts the installation process EI IL Gallito 2 0 6461 12 being installed Please wall 2 2 5 Step five End of the installation This screen lets the user know that product installation is complete By clicking on FINISH you will access the last screen in the installation process 2 0 b4Bits has been successfully installed Click Cloze to exit Please use Windows Update to check for any critical updates to the Framework Cancel lt Back 3 SUMMARY OF POSSIBLE ACTIONS The application makes it possible To generate semantic spaces under various parameters To find the similarity between terms documents and pseudodocuments e To generate lists of neighbors under various parameters with the possibility of exporting them to Microsoft Excel To load and save semantic spaces in different formats e To generate txt files from significant matrices To batch process ne
Download Pdf Manuals
Related Search