Galileo*CATPAC: - The Galileo Company
Contents
1. CATPAC OUTPUTS ci wetter teat Sila RESEARCHERS da bs pr etai ge doce Kee 17 R Summary ceperint SADA acetate ajad dba Cere 17 Dendog ami eR e ladies bn E wate Ea SRI TERR DRY e 18 Weight input networks WIN file r e ee 18 Coordinates the CRD file 19 SOME FINAL REMARKS n een rererere 20 Appendix 1 Tools asuksid A Med S WU ees E ee EVER URL Ue 21 vii CATPAC Terra Research viii Terra Research INTRODUCTION TO CATPAC CATPAC is a self organizing Artificial Neural Network that has been optimized for reading text CATPAC is able to identify the mostimportant words in a text and determine their patterns of similarity based on their associations in the text From this information it is able to tell you the main concepts dealt with in the text CATPAC does this by assigning a neuron to each major word in the text It then runs a scanning window through the text The neuron representing a word becomes active when that word appears in the window and remains active as long as the word remains in the window Up to N words can be in the window at once where N is a parameter set by the user As in the human brain the connections between neurons that are simultaneously active are strengthened following the law of classical conditioning The pattern of weights or connections among neurons2. CRD extension afterthe F1 prompt This CRD file provides the basis fora wide variety of analysis of CATPAC data including perceptual maps development of marketing and advertising strategies and tracking of perceptual change Figure 4 shows how the same clusters portrayed in the dendogram above this time in the form of a perceptual map To make the plot more readable we used an option in PLOT that allows the user to remove concepts from the display In this case we retain only the words in the major clusters Notice at the top of the plot the three concepts DOMINO S FAST and DELIVERY To the right and lower LITTLE CEASER INEXPENSIVE TDOMINO IDIZIVIRY PI224 BUT i UUALLTY Figure 4 Galileo Map of Pizza Interviews ONE and TWO Right in the center are the terms PIZZAHUT QUALITY and LIKE 18 Terra Research T h i s perceptualmap provides not only an alternative way to represent the same results as the dendogram but it allows for a wide A range of TAISAR special LITTLE analyses INEXPENSIVE DELIVERY Figure 5 shows TUO the same perceptual map i n 2 dimensions A WU ALLY CUNE For a detailed description of these features refer to the Galileo PLOT manual provided with your diskette SOME FINAL Figure 5 2 dimensional map of pizza interviews REMARKS CATPAC represents a new generation of artificial neural software that can do things older
3. and third order relationships among the words to be considered Very little cycling or none at all as in the simple co occurrence model tends to find only very superficial associations Too much thinking however is not always a good thing since CATPAC can tend to see things as all pretty much alike if its allowed to cycle too many times Some analysts with a warped sense of humor like to refer to this problem as The Buddhist Monk syndrome since after sufficient contemplation it appears that all things are one Clamping When a word is found in the window its neuron is activated But it can become de activated again as the network goes through its normal processes just as you yourself see things become aware of them and then forget them If you never forgot your mind would become so cluttered with images in only a few minutes that you could not go on with life When you choose to clamp the nodes another word forneuron you prevent them from turning off again It s like writing yourself a note and holding it in front ofyou so you mustalways pay attention to the words in the note Zelf analysis A self analysis will allow you to identify self references based on predetermined id s within your data set These points will be plotted as cloud using plot This option allows you to determine the rough boundary and location of the self point of the text There are two ways to do a self analysis For either you have to include as
4. computer software couldn t In this manual we ve tried to acquaint you with some of the new possibilities this technology makes available But neural technology is so new that not even the development community has a good understanding of what s possible yet Your best strategy is to spend time with the program and experiment If you have any problems please call your Terra representative Appendix 1 Tools YourGalileoinstallationincludes a directory called GALILEO TOOLS On this directory Terra has supplied three helpful DOS tools Firstis a simple read only editor called LOOK LOOKis a public domain program which allows you to examine the contents of any file interactively It is convenient since you can page up and down or scroll up down left and right in the file using the cursor control keys You can also easily read the 132 column format files that V55 writes And since LOOK is a read only editor you don t run the risk of altering important files To use LOOK simply enter the command LOOK filename at the DOS prompt To leave LOOK press ESC 19 CATPAC Terra Research Also included is a very powerful ASCII editor EDWIN EDWIN is a public domain program which follows the formats of WORDSTAR and can be very helpful in modifying files produced by V55 foruse in the other Galileo programs and vice versa EDWIN has complete online help accessed by pressing F2 once in the program To start EDWIN simply enter the command EDWIN
5. forms a representation within CATPAC of the associations among the words in the text This pattern of weights represents complete information about the similarities among all the words in the text Technically the pattern of connections among neurons is a complete paired comparison similarities matrix and so lends itself to the most powerfuland sophisticated of statistical analyses Among these is the diameter method cluster analysis automatically performed by CATPAC CATPAC can automatically exclude from consideration any arbitrary list of words A default list of articles prepositions and the like is contained in a file labeled EXCLUDE DAT You can add ordelete words to this file by using the enclosed program EXCLUDE EXE CATPAC expects to find both EXCLUDE DAT and EXCLUDE EXE in a directory called C GALILEO RUNNER This directory is automatically created during installation If you have not followed the standard installation instructions you mustcreate such a directory and copy all files with an EXE extension as well as the EXCLUDE DAT file to this directory You can create you own exclude files and we recommend that you do so Every data set is different and there are some words you may wish to exclude in one but not another To use an exclude file you have created simply enter it s name in the Exclude file field We also recommend that you place the directory C GALILEO RUNNER in yourpath This will allow you to run CATPAC from any di
6. in yourcurrent directory be sure to specify its exact path Make sure you Exclude file is in ASCII and has only one work on each line Include file Justas you can exclude certain words froman analysis you can include others After stripping some words you may wish to analyze may be dropped fromthe data set If you wish to analyze these words anyway place themin an ASCII file in the same format as your Exclude file and enter the file name here A second use for the Include file is for a Zelf analysis named after the famous Rudolph Zelf from Vienna If you choose to do a Zelf analysis the Include file contains words that may reference respondents in the text E G I Me MY aname orany specific identifieryou used in you data collection do not specify your Id number in this file CATPAC will associate the Id number with the self references automatically File descriptor This prompt is asking you for the filename you want CATPAC to use to as a prefix when labelling your output files You may use any name you wish up to 8 characters When you are doing a network run CATPAC willautomatically create several files using the filename you entered and supply an extension for each a labels file LBL an output file CAT a weighted input network file WIN and a coordinate file CRD The labels file contains the unique words CATPAC identified in the analysis while the output file contains the basic CATPAC outputs i e the word counts
7. suggest you use the CAT extension If you leave this field blank CATPAC will not run and will prompt you for a file name 12 Terra Research FREOUENCY ANALYSIS CATPAC has the capability to read every word in your file and list these words in descending order of freguency as well as in alphabetical order A FREOUENCY RUN is typically used by analysts to help them clean their data prior to performing advanced analyses A FREOUENCY RUN can help you find typographical errors synonyms plurals pro nouns and other such words that you may want to recode using a word processing program or a text editor prior to proceeding with further analyses Ifyou want to perform a Freguency analysis type F1 at the main menu Here is the screen you ll see This is a very simple screen Most of the information PFREOUTCNYTILCS reguested is Run title explained X ON uuludu lile osa vu utervamuludu dal O t lt Huipul fe ou mus supply an output file name for your frequencies Cursor to the Output file field and enter a file name to store this Is to exit Hiire mm The job output on he Mesa SAHA oru OS A otiam ae one 6 Frequency analysis screen 13 CATPAC Terra Research HELP You may get help at any time by pressing F3 There are two ways to get help If you press F3 from the main menu you will be given the option to press any other function key for help on any of CATPAC s main operations Whe
8. vary between 1 is 0 0 and 5 for the logistic varying between 0 and 1 By lowering the threshold you make it more likely for neurons to become activated by raising the threshold you make it less likely for neurons to become activated Decay rate When you see an object neurons which represent that object are activated When the object is gone CATPAC Terra Research the neurons fortunately turn off again If they didn t you d be seeing everything you ever saw all the time The decay rate specifies how guickly the neurons return to their rest condition 0 0 after being activated The default rate is 9 which means that each neuron if not reactivated will lose 90 of its activation each cycle Raising the rate makes them turn off faster lowering the rate means they are likely to stay on longer Learning rate When neurons behave similarly the strength of the connection between them is strengthened The learning rate is how much they are strengthened in each cycle Default is 001 Increasing this rate makes CATPAC learn faster Faster is not always better though since too high of a rate can make CATPAC oscillate back and forth as new information is read No one knows the optimumrate oreven if there is an optimum rate so feel free to experiment Output files As explained above CATPAC automatically produces its out files fora Network analysis These four fields Catpac file Label file Win file and Coordinate file merely dis
9. 20 3 Occurrence They show GOOD 7 6 1 28 20 3 INEXPENSIVE 9 7 8 37 26 8 that Little was the FAST 7 6 1 26 18 8 LIKE 6 5 2 21 15 2 f 1 LIKE 6 5 2 21 15 2 LITTLE 13 11 3 58 42 0 most reguently DELIVERY 6 5 2 26 18 8 ONE 6 5 2 2719 6 occurring word that it YOU 6 5 2 26 18 8 PIZZA 3 2 6 12 8 7 ae i ONE 6 5 2 27 19 6 PIZZAHUT 7 6 1 3525 4 occurred 13 times which QUALITY 4 3 5 20 14 5 PLACE 3 2 6 15 10 9 WANT 4 3 5 20 14 5 QUALITY 4 3 5 2014 5 Was 11 3 of all PIZZA 3 2 6 12 8 7 TWO 7 6 1 32 23 2 occurrences Little FASTER 3 2 6 11 8 0 WANT 4 3 5 20 14 5 d in 58 or 42 0 PLACE 3 2 6 15 10 9 YOU 6 5 2 26 18 8 appeared or 22 0 70 of the scanned Figure 2 CATPAC WORD COUNTS windows This last figure is referred to in the output as a CASE FREQ and indicates the number of times a given word appears in a case If you had delimited each response with a 1 and done a case by caseanalysis The CASE FREQ would indicate the number of respondents to mention each word The words pizza faster and place occurred least often three times each CATPAC didn t considerany words that occurred fewer than three times since that would have resulted in the identification of more than the 20 unique words we requested The right most columns give exactly the same information as the left mostcolumns exceptthe unique words are now listed in alphabetical order for easy look up Dendogram Figure 3 shows the output from the hierarchical cluster analysis These p
10. 23A Durham Drive Amherst NY 14228 Galileo CATPAC User Manual and Tutorial Rev June 1993 5 0 CATPAC Terra Research CATPAC copyricut 199 By J osten VV ALL RIGHTS RESERVED NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS ELECTRONIC OR MECHANICAL INCLUDINGPHOTOCOPY RECORDINGOR ANY INFORMATION STORAGEAND RETRIEVAL SYSTEM WITHOUT PERMISSION IN WRITING FROM The Galileo Company CATPAC GALILEO and ORESME are trademarks of The Galileo Company All other brand and product names are trademarks or registered trademarks of their respective companies PLEASEDON T LOSETHIS PAGE IT CONTAINS THE REGISTRATION NUMBER YOU WILL NEED TO UPGRADE TO LATER RELEASES OF CATPAC Your Name Your Registration Number Version 3 0 Copyright 1990 The Galileo Company All Rights Reserved ii Terra Research IMPORTANT PLEASE READ CAREFULLY BEFORE USING THE SOFTWARE NOTIFICATION OF COPYRIGHT THIS SOFTWARE IS A PROPRIETARY PRODUCT OF The Galileo Company AND IS PROTECTED BY COPYRIGHT LAWS AND INTERNATIONAL TREATY YOU MAY MAKE A REASONABLE NUMBER OF COPIES OF THIS PROGRAM FOR BACKUP PURPOSES AND YOU MAY COPY THE SOFTWARE TO THE HARD DISK OF A SINGLE COMPUTING PLATFORM OF THE TYPE SPECIFIED IN YOUR LICENSE YOU ARE PROHIBITED FROM MAKING ANY OTHER COPIES OF THE SOFTWARE FOR ANY OTHER PURPOSE BY COPYRIGHT LAWS YOU MAY MAKE ONE COPY OF THE WRITTEN MATERIALS ACCOMPA
11. NYING THIS SOFTWARE FOR ARCHIVAL PURPOSES The Galileo Company PLEASEREADTHISLICENSEAGREEMENTBEFOREUSINGTHESOFTW ARE THIS AGREEMENT IS A LEGAL CONTRACT BETWEEN YOU AND The Galileo Company GOVERNING YOUR USE OF THIS SOFTWARE USING THIS SOFTWARE INDICATES YOUR ACCEPTANCE OF THIS AGREEMENT IF YOU DO NOT WISH TO ACCEPT THETERMS OF THIS AGREEMENT PLEASE RETURN THE UNOPENED SOFTWARE PROMPTLY TO The Galileo Company IF YOU HAVE ANY QUESTIONS ABOUT THIS AGREEMENT PLEASE CONTACT The Galileo Company 23A Durham Drive Amherst NY 14228 TERMS OF LICENSE THIS IS AN EXPERIMENTAL PROGRAM WHILE The Galileo Company CERTIFIES THAT THE HIGHEST STANDARDS OF DILIGENCE AND SCIENTIFIC INTEGRITY HAVE BEEN APPLIED TO THE DEVELOPMENT OF THIS SOFTWARE BY ACCEPTING THIS LICENSE YOU AGREE THAT THIS IS EXPERIMENTAL SOFTWARE AT THE CUTTING EDGE OF SCIENTIFIC PROGRESS NOTASMUCHISKNOWNABOUTTHEPERFORMANCE OF NEURALNETWORK TECHNOLOGY ASISKNOWN ABOUT TRADITIONALCOMPUTER SOFTWARE YOU AS THEEND USER AGREE THAT REASONABLE AND PRUDENTCAUTION ABOUT THEAPPLICATION OFRESULTS FROM THIS SOFTWARE IS APPROPRIATE AND The Galileo Company AGREES TO SHARE WITH YOU THE LICENSEE RELIABLE ESTIMATES OF THE OPERATING PARAMETERS OF THE SOFTWARE INSOFAR AS THEY ARE KNOWN BY TERRA lii CATPAC Terra Research The Galileo Company GRANTS YOU THE RIGHT TO USE ONE COPY OF THE SOFTWARE ON A SINGLE USER COMPUTER EACH WORKSTATION OR TERMINAL ON A MULTI USER COMPUTER SYSTEM OR LOCAL AREA NE
12. TWORK MUST BE LICENSED SEPARATELY BY TERRA RESEARCH AND COMPUTING COMPANY YOU MAY NOT SUBLICENSE RENT OR LEASE THE SOFTWARE TO ANY OTHER PARTY YOUMAYMAKEREASONABLE BACKUP OR ARCHIVALCOPIES IFTHESOFTWARE BUT YOU MA YNOTDISASSEMBLE DECOMPILE COPY TRANSFER REVERSEENGINEER OR OTHERWISE USE THE SOFTWARE EXCEPT AS STATED IN THIS AGREEMENT LIMITED WARRANTY The Galileo Company will replace defective diskettes that are returned within 90 days of the original purchase date without charge The Galileo Company warrants that the software will perform substantially as stated in the accompanying written materials If you should discoverany significant defect and report it to The Galileo Company within 90 days of purchase and Terra is unable to correct it within 90 days of receipt of your report of the defect you may return the software and Terra will refund the price of purchase SUCH WARRANTIES ARE IN LIEU OF OTHER WARRANTIES EXPRESSED OR IMPLIED INCLUDING BUT NOTLIMITED TO THEIMPLIED WARRANTIES OF MERCHANTABILITYAND FITNESS FOR A PARTICULAR PURPOSE WITH RESPECT TO THE SOFTWARE AND THE ACCOMPANYING WRITTEN MATERIALS IN NO EVENT WILL The Galileo Company BELIABLE TO YOU FOR DAMAGES INCLUDING ANY LOSS OF PROFITS LOST SAVINGS OR OTHER INCIDENTAL OR CONSEQUENTIALDAMAGES ARISINGOUT OF YOUR USEOF OR INABILITY TO USE THE PROGRAM EVEN IF The Galileo Company OR AN AUTHORIZED TERRA REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES
13. The Galileo Company WILL NOT BE LIABLE FOR ANY SUCH CLAIM BY ANY OTHER PARTY This limited warranty gives you specific legalrights Some states provide otherrights and some states do not allow limiting implied warranties or limiting liability for incidental or consequential damages Forthis reason the above limitations and or exclusions may not apply to you Ifany provision of this agreement shall be unlawful void or for any reason unenforceable then that provision shall be deemed separable from this agreement and shall not affect the validity and enforceability of the remaining provisions of this agreement This agreement is governed by the laws of the State of New iv Terra Research York CATPAC Terra Research U S Government Restricted Rights The software and accompanying materials are provided with Restricted Rights Use duplication or disclosure by the government is subject to restrictions as set forth in subparagraph b 3 ii of the Rights in Technical Data and Computer Software clause at 252 277 7013 Contractor manufactureris The Galileo Company 23A Durham Drive Amherst NY 14228 vi Terra Research TABLE OF CONTENTS INTRODUCTION TO CATPAC ic oia A GLE RUIN de tad 1 INSTALLING CATPAC oii 10454 bee cae eee cage Hee a da oeh teg epi b 2 RUNNING nen bo A baie dn VEDI EE edie ile PED 2 SCREEN CONTROL td dd be e Ghats al 4 NETWORK ANALYSIS tiet as RAE Wem UP AALL SA et
14. aa 5 ee eek opc via pO kle ah Sii AOE Sw sarvedeta MER 5 Data fil s cae wae ke ea hh expel Sad wo ee ep MEE E 6 Exclude filene o eI PER Sieg ies Re ag eg tend 6 Include file Begone sited bah atu BSA AAR REN BR eh ly A AEE TS 6 File descriptor consis coals ie eu SWR EE BE ER OE Roh e ERE A EES 6 Unique words Ma od sous Miata os ea natn use ph elas sepia aidan de ad vind eid dn 7 Window size 0r L Rr 8 Slide size AA Sipa AA A SA ee alata ES 8 Cycles EE 8 Clamping e pec A ERR IR UE eee qe E E een 9 Zelf an lysis A Ad cl gate tay ans Gs Ee NDA pad DEPO Ras 9 Network parameters DU M RUM PEOR RD CEN ae Caves pts 10 Function TO ML cree Euer e es SRI DAS RSS 10 Threshold via eL AS ae ORE te Heed Cee See y Meee alah 10 Decay Tale oss oso tess tpetrus dade hae Carte bias patustas ale 11 Learnitig TALC ii ia A Vet HAE T deese 11 Output files ia ENS PU EROR A uie eR UTE 11 COMPLETED SCREEN eR x 12 CLUSTER ANALYSIS v5 ey tl cents a be eee COS LR ie Se ae 13 Matrix E ahs Heke iO etu eT daa 13 Catpac file o eU A A naat BS ae AJA ege 13 FREQUENCY ANALYSIS eee MERGER A RIEF DIC la 14 HEEP dla tala nc sto este d i M ahud 15 RESET TO DEFAULTS uy ERR RYE VERA Werbe ae 15 INPUT TO CATPAC AA Ait a REIS 16
15. alphabetic listing and clusteranalysis The weight input network contains the connection strengths between nodes and the coordinate file is used by PLOT to make a perceptual map It contains the coordinate loadings ofeach unique word on a 3 dimensional set of axes The more ambitious can actually draw their own map by hand using the x y and z coordinates listed in this file For example if you were analyzing a text file that contained information about cars you might use the filename CARS and CATPAC would create the following files CARS LBL and CARS CAT CARS WIN and CARS CRD These file names are displayed automatically at the bottom of the Network analysis screen Unique words At this field CATPAC is asking youhow many words you want to carefully study Most of the time you will only want to use only the top fifteen or twenty or 30 unique words This version of CATPAC can perform higher order analyses on as many as 150 words If you need to study more than 150 words call TERRA we have another version of the program that can read more words but you will Terra Research need special instructions and perhaps a faster machine CATPAC identifies unique words in the following manner First the programlooks at every word that occurs once or more and then checks to see if the number of such words is greater than the number ofunigue words you specified If it is CATPAC will study every word that occurs twice or more and then check t
16. among a set of neurons which stand forthe main words in the text This network of interconnected neurons is the main product of CATPAC s thinking and itis stored in a file with the suffix WIN This matrix is not of much use by itself but it contains all the information CATPAC has learned about the text and can serve as an input to other programs including ORESME ORESME simulates the processes a neural network carries out when it receives inputs of various kinds of information and can give an interesting simulation of how individuals or groups of people might respond to different messages Since the WIN matrixis in a standard ASCII format with its format statement in front ofthe data it may be input into a variety of standard statistical and mathematical analyses software packages Coordinates the CRD file 17 CATPAC Terra Research When you reguest a network analysis CATPAC will also produce a file of spatial coordinates which has the same generic name as the other files produced by CATPAC but which ends in the extension CRD This file contains information which can be used to generate a pictorial representation of the word associations CATPAC discovered during its analysis of the text The CRD file contains the coordinates of the words on the basis of which plots can be made To make a perceptual map the user mustcall up the program PLOT which is also provided with your software and type this file name with the
17. at the DOS prompt You can also enter a file directly with EDWIN by entering the command EDWIN filename If you already have an ASCII editoryou favor you may use that instead of EDWIN For more information on installing and using EDWIN consult the documentation provided on the GALILEO TOOLS directory The last tool provided is called UP Up lets you climb up your directory tree in only three keystrokes If your default directory for example is GALILEO DATA then issuing the command UP at the DOS prompt will set your default directory to GALILEO Issuing the command again will move you to the root directory All three ofthese utilities are public domain software and are neitherwarranted nor supported by The Galileo Company The Galileo Company and Computing or any of their agents The are provided at no charge as a convenience for the user Note that authors of public domain software sometimes request voluntary payments from users for the use of their programs No such payments have been made on your behalf by Terra Galileo or any of their representatives 20
18. ber as the number of unique words you wish to study Window size or 1 CATPAC works by passing a moving window of size n through your file If you were to enter a window size of 7 a good guess to start with in most cases CATPAC would read your text seven words at a time So for example if you were to specify a window size of 7 and a slide size of 1 CATPAC would read words 1 through 7 then words 2 through 8 then words 3 through 9 and so on Any time a word is in the window the neuron representing this word becomes active Connections among active neurons are strengthened so words that occur close to each other in the text tend to become associated in CATPAC s memory If you enter 1 instead CATPAC abandons the moving window and looks for 1 s in your data file These mustoccurin columns and 2 and all the text that lies between these delimiters is considered a case If you use the moving window model you do not need any 1 s in your file and CATPAC will make its own cases automatically using the window size you specified Further having 1 s in your data file will not adversely effect your run if your choose to read your file using a moving window Slide size This prompt is asking you how you would like the moving window to slide through the text The number you select dictates how many words the window will skip prior to reading the text You may select any increment you like For example if you chose a windowof5 and a
19. ictures are called dendograms and they look a bit like the skyline of a city seen from afar The buildings underneath the words show which words cluster together 16 Terra Research As can be seen in Figure 3 this cluster analysis reflects the information contained in the text guite well The words little and caesar clustervery sharply togetheras we would expect Domino and Fast cluster very closely and the word delivery joins this cluster at a slightly lower level on Domino s specializes in fast delivery Similarly Pizzahut quality and like form a third cluster As we move downward PIZZA CLUSTER ANALYSIS through the diagram each of the clusters DIAMETER METHOD grows larger including more and more terms GPDFDYWFLPQTOLCIP OIOAEOAAIIUWNIANL Eventually the first OZMSLUNSKZAOETEEA cluster includes DZ I T I TIE is TSX C AI LAPE Domino ast Ot EET EP delivery you along NE S with the subcluster ES i m you want and su s faster Little Caesar ends up in a cluster which includes little DIE DAMM ll caesar inexpensive eee and place with the subcluster two and one Little Ceaser offers two for one pizzas AAR AAA Weight input networks cae RRNA RAMA WIN file As we ve said works by finding the Figure 3 CLUSTER ANALYSIS DENDOGRAM connections
20. n you are done with help on the main menu press F7 as the screen indicates If you are filling out any specific analysis screen press F3 on any field and you will get a brief explanation of what the field is and what kind of information it is looking for RESET TO DEFAULTS Ifyou have altered any of CATPAC s default settings and wish to return to these values press F5 and CATPAC will reset all changed parameters to the ones supplied with the program 14 Terra Research INPUT TO CATPAC CATPAC can read any text file that has been converted to ASCII Some examples of text files people have studied using CATPAC include answers toopen endsurvey questions focus group transcripts newspaper and magazine articles down loaded from a data base comments left on a customer telephone hot line and restaurant hotel airline comment cards When preparing data for a CATPAC analysis keep in mind that while CATPAC is guite an amazing technology it is still quite primitive In as much if you are studying focus group transcripts for example you should probably firstparse the file into discrete topic specific sections ratherthan have CATPAC try to study a file that spans 10 15 topics Figure 1 shows a text derived from some interviews where people were asked to describe the difference between a select set of pizza restaurants Asking people todescribe the difference between products is usually a good method since they then usually report attrib
21. o see if the number of such words is still greater than the number you requested If it is CATPAC will study thosethat occurthree or more times and so on until it finally obtains the number of unique words you specified Many times CATPAC will provide you with fewer unique words than the number requested When this happens it means that there were several words which occurred with the same frequency as the nth unique word and if included the number of unique words identified would have exceeded the number you requested Hence to avoid giving you more than what you ask for CATPAC deletes all of these words leaving you with fewer unique words A very good reason to conduct a FREQUENCY RUN see below priorto doing any other analysis is to examine the frequency with which the words you wish to study occur in the text Doing this will help you determine exactly how many words you should specify at this prompt The process is simple Using the descending frequency list on your initial FREQUENCY RUN examine the rank order position of all words you want to include in your analysis Find the rank order position of the last word you want to include in the analysis and specify that as the number of unique words you want CATPAC to study Remember if there are other words which occur with the same frequency as the last word you want to include you mustcount down to the rank order position of the last word with the same frequency and specify that num
22. play the names of these files based on the File descriptor you entered These fields can only be altered by changing the contents of the File descriptor field 10 Terra Research COMPLETED SCREEN A completed Network analysis screen might look like this KT VOA ayal RIS 4 Completed Network screen 11 CATPAC Terra Research CLUSTER ANALYSIS To do a Cluster analysis press F2 at the Main Menu The cluster analysis will appear Many of the fields for a Cluster analysis identical to Run tile those for a a B Network analysis The m a i n differences are lack of network Pilari rires parameters and how you enter the output file names CLUSTER AK AL Y Incl Tigu Matrix file Cursor down to the Matrix file field and enter the name of a file to store your co occurrences If you enter a file name at this field CATPAC will output a file that contains a listof word co occurrences it encountered within the window size you specified CATPAC will list every co occurrence and tell you how many times it encountered each co occurrence If you leave this field blank CATPAC will not produce this output I to exit 111 nm Tha jele he Hesa rae TT Read RL 5 Cluster analysis screen Catpac file You must entera file name for this field This is the same output mentioned above CAT and is the standard output for CATPAC You may call it anything you like but we
23. rectory simplify your analysis set ups and keep your raw data and or results separate from your program files If you are not sure how to edit the path statement contained in your AUTOEXEC BAT file consult your DOS manual CATPAC Terra Research INSTALLING CATPAC Place the diskette in the A or B drive Type INSTALL lt diskette drive gt lt target drive gt and press Enter For example to install the system on your C drive with the diskette in the A drive you would type INSTALL A C That s it The install program will take care of everything The following directories will be created GALILEO RUNNER Contains the executable programs GALILEO HELP Contains the help files GALILEO DOC Contains all available Galileo Documentation in WordPerfect 5 0 format GALILEO DATA Contains sample data sets GALILEO TOOLS Contains a text editor and several utility programs RUNNING CATPAC You will need at least530k of free ram to run Catpac If you experience difficulty running the program chances are you will have to free up some memory The best way to rectify a memory problem if you have a 386 or better processor is to install a memory manager like QEMM This program is inexpensive and will give you up to 630k of ram to run programs with Another approach would be to create a vanilla boot diskette that has the bare minimum of TSR s and device drivers in Autoexec bat and Config sys If you are unfamiliar with these files or are
24. slide size of 1 CATPAC would read words 1 through 5 2 through 6 etc If you chose a windowof5 and a slide of 2 CATPAC would read words 1 through 5 then 3 through 7 etc Slide sizes largerthan 1 are mostoften used when CATPAC Terra Research you have a very large text file from which you want to draw samples This is a new field so feel free to experiment If you entered 1 for window size case by case analysis this field is ignored Cycles CATPAC s network analysis procedure works in the following manner When words are present in the scanning window theneurons assigned to those words are active and the connection among all active neurons is strengthened But the activation of any neuron travels along the pathways orconnections among neurons and can in turn activate still otherneurons whose associated words may not be in the window These neurons can in turn activate still otherneurons and so on In an actual biological neural network these processes go on in parallel and in realtime so that the signal coming into the network is spreading at different rates of speed throughout the network and neurons are becoming active and inactive at different times This process of delay is called hysteresis In a serial computer like yours however this is extremely difficult process to model and so the network is updated periodically all at once Each update is called a cycle Letting CATPAC cycle two or three times allows second
25. the firstline of each case an id number that begins with either or Method I locate by id simply associates the Id number Terra Research ofany given case with the case Method Il locate by self reference will essentially replace any self referent referred to in your Include file with the id and treat that id as a node Since method II forces a repetition of the id this will lead to stronger associations between the self point and the text if the text is sufficiently rich in self reference If there are no or few self references the Id method ought to be used To do a self analysis press Enter on the self analysis field You will be given a pop up menu of choices Using the arrow keys choose the method you wish and press Enter again Remember if you choose to do a self analysis your Include file becomes a self reference file Network parameters CATPAC can simulate four different kinds of neurons functional forms and the overall performance of CATPAC depends onthree parameters threshold decay rate and learning rate The most generally useful neuron and some reasonable values for the three general parameters have been chosen as defaults in CATPAC But you can change them if you wish and none of these neuron types or parameters are sacred even those selected by Terra as defaults You might well find CATPAC performs better for some tasks with a different choice of neurons and or default parameters In order to change any defaul
26. the job If you wish to bail out and go back to the Main Menu press F7 One nice feature of CATPAC is the ability to do multiple runs with the same settings If you decide to analyze 10 data sets each with the same settings you need enter the settings but once You need only change the Data file and File descriptor field or output file fields for a Freguency or Cluster analysis Terra Research NETWORK ANALYSIS This is the premier analysis offered by CATPAC accept no substitutes If you want to perform a Clusteranalysis based upon simple word co occurrences you would have typed F2 Essentially this was way allcluster analyses were performed before the development of artificial neural networks like CATPAC You however choose to conduct a more advanced form of relational analysis a Network Analysis so you pressed F1 In a Network analysis CATPAC will generate square matrix of numbers which summarizes the connection strength between unique words How catpac performs this analysis is discussed in detail below as we explain the fields specific to a Network analysis Be aware that if you want to produce a perceptual or brand map you must choose a Network analysis If you Press Fl to do a Network analysis you will see the following screen Some of the fields for a Network analysis are also offered for an dices au es eas a a Cluster analysis and a Frequency run so we ll explain them here Many of the fields are supplied
27. ts just tab to the field of choice and enter a different value Function form This option allows you to try different transfer functions A true chiphead would jump at the chance to play with these You can choose from four a logistic varying between 0 and 1 a logistic varying between 1 and 1 a hyperbolic tangent function varying between 1 and 1 and a linear function varying between 1 and 1 Some writers speculate that different functions are better for different kinds of task but no one knows for sure at this time The default threshold is 0 0 If you choose the logistic function that varies between 0 and 1 the threshold will automatically be set to 5 If you d like to experiment with different transfer functions press enter at this field and you will get a menu of the fourforms arrowto the function of your choice and press enter Note A Chiphead is a person with an exceptional commitment to computing If you plan to do basic research on various transfer functions you are one Threshold Each neuron in CATPAC is either turned on by being in the moving window or else receives inputs from other neurons to which it is connected These inputs are transformed by a transfer function Afterthe inputs to any neuron have been transformed by the transfer function they are summed and if they exceed a given threshold that neuron is activated otherwise it remains inactive The default threshold for the three transfer functions that
28. unsure how to make a boot diskette refer to your DOS manual or resident Techie There are 3 ways to access CATPAC 1 Type the word GALILEO and a menu will appear on the screen You would then select the number which corresponds to CATPAC If you have not edited your path to include GALILEO RUNNER you must first change to C GALILEO RUNNER prior to typing GALILEO 2 If you have placed C GALILEO RUNNER in your path you can simply type CATPAC 3 If you have not edited your path you must first change to the directory C GALILEO RUNNER and then type CATPAC If you have not installed a memory manager or have a 286 machine and are having difficulty running the program try either method 2 or 3 These methods require less memory because they do not use the menu interface Terra Research Once you call up the program CATPAC will display a menu screen like the one below To make a selection press the function key t h a t corresponds to the operation you wish to preform You can Freyusi eiss choose one of 5 Resol bv deluulta three types of analysis Help or Reset to defaults Each of these options is explained below b Mesa iII yyh ANT A IS A 2 CATPAC main menu CATPAC Terra Research SCREEN CONTROL Each analysis screen has a numberof fields that require information to run see figure 3 You move from field to field using the arrow or tab keys Once you have entered all the information press F10 to run
29. utes which make a difference instead of attributes which all the products might share The reader will note that this particular text file is not very long or of very high quality Hopefully your data set will be a little better I like pizza hot and fresh I like quick delivery like Domino s gives but I need quality like pizzahut Little Caesar s is inexpensive but I guess pizzahut has quality Domino s delivers but Domino s is expensive Little Caesar s is inexpensive and you get two at Little Caesar s Little Caesar s two for one deal is inexpensive I like good flavor like pizzahut but I guess Domino s is faster Sometimes you want it faster and Domino s is faster If you want good flavor Pizzahut is for you but if you want it inexpensive Little Caesar s is the best It s good Little Caesar s is good but Pizza Hut is good too Domino s is not as good but fast Domino s is fast I think Domino s has fast delivery and Domino s fast delivery means a lot to me Pizzahut s quality is important but it s not worth it Little Caesar s two for one is really good Two for one Little Caesar s is the two for one place Pizzahut quality sets it apart but Little Caesar s is inexpensive Pizzahut is expensive But of course Domino s fast delivery can be important When you want fast delivery Domino s is the fast delivery place For inexpensive pizza Little Caesar s is most inexpensive of all Inexpensive little caesar s is the place for t
30. with default values and file names that we have found to work best with a wide variety of runs You are 3 CATPAC Network analysis screen encouraged to change themat your whim to experiment with different settings Every data set is different and what works for one may not work for another KT VRDL ayal YSIS Run title CATPAC is asking you for a title which will be printed as a banner on your output file Your title is limited to forty characters including spaces Data file CATPAC needs to know the exact path to your data You must specify the drive the directory and the name of the file which contains the text you wish to analyze If you are running CATPAC fromthe directory where your data are stored you can simply type the name of the file which contains your CATPAC Terra Research data Remember CATPAC can only read an ASCII file Sometimes people make the error of exiting their word processor without first converting the file to ASCII Don t make this mistake We like to use the extension TXT to denote an ASCII file That way when we see this extension on a file we know CATPAC can read it Exclude file CATPAC will let you use any EXCLUDE file you wish If the one we sent you EXCLUDE DAT meets yourneeds simply skip this field to accept the default EXCLUDEfile Alternatively if you want to use another EXCLUDE file you have created for a special purpose type the name of this file Again if this file is not
31. wo for one little caesar s two for one Little Caesar s is inexpensive Figure 1 PIZZA INTERVIEWS 15 CATPAC Terra Research CATPAC OUTPUTS Getting CATPAC to analyze these interviews is very simple In this case we asked CATPAC to cycle once and to identify no more than 20 unique words We set the window size to 5 no other values were re set The results are shown in Figures 2 and 3 Run Summary Figure 2 shows the most basic output of CATPAC It consists of CATPAC PC v3 00 a summary of the 05 25 93 09 29 40 parameters selected and a freguency count ofthe TITLE Pizza interviews main words found in the DATA FILE PIZZA TXT text It shows that there TOTAL WORDS 115 THRESHOLD 000 were 115 total words in TOTAL UNIQUE WORDS 17 RESTORING FORCE 100 the text and that 17 TOTAL WINDOWS 138 CYCLES 1 TOTAL LINES 21 FUNCTION Sigmoid 1 1 unique words were WINDOW SIZE 5 CLAMPING Yes SDE found There were 138 windows inthe analysis DESCENDING FREOUENCY LIST ALPHABETICALLY SORTED LIST and 21 lines of text CASE CASE CASE CASE WORD FREO PCNT FREO PCNT WORD FREQ PCNT FREQ PcNT The left most columns LITTLE 13 11 3 58 42 0 CAESAR 13 11 3 57 41 3 Present the major words CAESAR 13 11 3 57 41 3 DELIVERY 6 5 2 26 18 8 in descending order of DOMINO 11 9 6 46 33 3 DOMINO 11 9 6 46 33 3 INEXPENSIVE 9 7 8 37 26 8 FAST 7 26 18 8 fre quency of PIZZAHUT 7 6 1 35 25 4 FASTER 3 2 6 11 8 0 TWO 7 6 1 32 23 2 GOOD 7 6 1 28
Download Pdf Manuals
Related Search