biolab / text-semantics

The package with scripts for semantic analyser project
MIT License
4 stars 5 forks source link

Finding interesting words with tfidf #14

Closed ajdapretnar closed 3 years ago

ajdapretnar commented 3 years ago

Script for finding interesting words with TF-IDF approach. Not bad, but not very exciting results either. Captures a lot of stylistic words.

PrimozGodec commented 3 years ago

It looks good. I just added two comments.

I think we should show results for the same documents in 02_05, 02_06, 02_07. Currently, 02_05, 02_06 show results for documents with index 4, 8 and this notebook shows results for 13. Since 4 and 8 are quite similar I suggest that we all show results for document 4 and 13. From this notebook 13 looks interesting.

ajdapretnar commented 3 years ago

Yes, I went with 13, because it was the most interesting. I can adjust the other notebooks to include the same documents.

PrimozGodec commented 3 years ago

I can go with 13 in my notebook since it is not merged yet.

PrimozGodec commented 3 years ago

@ajdapretnar can you just run the last cell such that results will be visible on the preview on Github. Or did you run it already and it is just not visible.

ajdapretnar commented 3 years ago

Fixed.

PrimozGodec commented 3 years ago

Thanks looks good