Saturday, 29 November 1997

ARDIS _ ARGIS

INPUTS
Yogesh/ cyril
·     Texts
    (A) Printed / Typed Documents

    (B) Electronic Documents

    Examples of Texts

(A)  Printed Documents
·     Bio - datas (Typed)
·     Job - Advts
·     Magazine / Newspaper Articles
·     Directories (Kothari /IRIS etc)
·     Directories (CA/ CS Membership)
·     Bullletins (Bapuji Impex/ Domex)
    (B) Electronic Documents
·     Databases
·     Bio - datas received over Internet / Extranet
·     E : mails
·     Files on Floppies
·     Files on CDs
·     Files on Hard - Disks
·     Files on Tapes
·     Voice / Speech Converted to text (thru speed - recognition software)

 PROCESS
·     Scanning of document (in a "scanner", if Printed/ typed document)
·     Electronic Scanning (if in electronic - file format)
·     OCR (for printed / typed document)
·     Spell - check / Automatic Spelling Correction
·     Search / Identify/ Pick - up "KEYWORDS" from each document / file
·     Give each document a unique "number" (this will be PEN for resume's / Advertisement Nos. etc)
·     Link key - words with document
·     Assign "meaning" to each keyword (based on context)
·     Store each keyword in relevant "Meaning - lines / Fields" to create a database”
·     For each keyword, create directory of "Synonyms & Autonyms"
·     Create a continuously up - date "tables" re : frequency - of - usage of each keyword
·     Create and continuously up - date "tables" re : frequency - of - usage of all words (not only keywords) used "before  & after" each keyword to create "CONTEXT PROBABILITY for each keyword.
·     Repeat above process for all the "Phrases" and "Sentences" in which a given KEYWORD has been used to establish "CONTEXT PROBABILITY" of phrase/ Sentences.
·     If a Keyword has been used in phrase / sentence having a very low "context Probability" then replaces that phrase/ sentence by a phrase / sentence of the highest "Context Probability".
 OUTPUTS :-
·     Databases of
·     Keywords
·     Synonyms
·     Antonyms
·     Bins/ Fields
·     Phrases / Sentences
·     "Frequency of Usage" tables
  PRINTED OUTPUTS:
·     Converted Bio - datas (Full)
·     Brief bio - datas (brief outline)
·     One - line tabulations
·     Standard Letters/ Responses
·     Short - lists of suitable executives.

h.c.parekh

No comments:

Post a Comment