search

SalikSayyed / Smartify

MIT License
1 stars 1 forks source link

HARDCODED GRAMMER ALGORITHM FOR BOOK TYPE.. (MAY BE MANUALLY PROVIDED) #18

Closed SalikSayyed closed 3 years ago

SalikSayyed commented 3 years ago

AS WE ALREADY SAW THERES NOT ENOUGH DATA TO TRAIN A MODEL TO EXTRACT KEYWORDS DIRECTLY FROM TEXT. SO ANOTHER WAY TO DO IT, IS BY CODING AN ALGORITHM (FINAL UNCHANGING HARDCODED) TO IDENTIFY KEYWORDS. Solution------ WE ALREADY SAW ITS POSSIBLE TO IDENTIFY TYPE OF WORD LIKE IF IT IS NOUN, PRONOUN, OR ANY PART OF SPEECH THROUGH TEXTBLOB LIBRARY. SO I THINK IF WE COULD TRY AND FIND A SPECIFIC PATTERN FOR KEYWORDS WE MIGHT BE ABLE TO ELIMINATE MANY 3 WORDS TUPLES

Like just saying for a Scientific book.. or a year book of some kind nouns can be an identifier for correct keywords.. We can then highligh. Like nobel prize winner somebody of year for somefield. Here somebody , year, somefield all are nouns.

So kind of like this..

Rahul214-Coder commented 3 years ago

before yesterday I am working on textblob for grammar ss mini3

SalikSayyed commented 3 years ago

okay. But as we discussed yesterday, There are better libraries than TextBlob so I think u are already working on TextRANK. Thats our only option bro For Final Text Map