Wordseer / wordseer

The WordSeer text analysis tool, written in Flask.
http://wordseer.berkeley.edu/
40 stars 16 forks source link

all "distinctiveness" scores seem to be 0 #230

Closed macfarlandian closed 9 years ago

macfarlandian commented 9 years ago

is this a preprocessor issue?

MartiHearst commented 9 years ago

Maybe this is a from __future__ import division issue?

macfarlandian commented 9 years ago

the JS app is requesting frequent words for the overlay menu from words_view.py ... it should be requesting them from associated_words_view.py.

THEN, associated_words_view.py should compute a tf*idf score for matching sentences within the slice (containing the co-occurring word) instead of just term frequency (as it currently does)

macfarlandian commented 9 years ago

pretty sure I reopened this because of the issues (mostly) fixed in db0ca46b5b11431d0637371ad1c943d8b05172f8, b89e2b3b1f4b603535acabebaf8c5ffc6ac177ab, f64309165660debebc5c08061df02744fe946b2a, so closing again

macfarlandian commented 9 years ago

see #244 for a lingering issue related to this