search

lbs-iitkgp / Opensoft-2019

0 stars 0 forks source link

Text rank, NLTK & weighing non-stopped words based on frequency in our corpus #43

Closed athityakumar closed 5 years ago

athityakumar commented 5 years ago

Gensim seems to capture the entire case & NLTK does decently too in capturing the facts - but the textrank is just not good for our case.

We can try tweaking nltk's scoring to be based on our legal corpus.

athityakumar commented 5 years ago

^ Sample files and output files can be seen in this branch: https://github.com/lbs-iitkgp/Opensoft-2019/tree/nltk-summarizer

athityakumar commented 5 years ago

Fixed by #77