Word Enrichment Analysis using VEctor Representations
Author: Kushal K Dey, Matthew Stephens
The main contents of this package are functions to do the following.
-
Robust word2vec - adaptive shrinkage analysis of word2vec cosine similarities to provide robust estimates
-
Network visualization - network visualization of the context of a word in a corpus based on word2vec ouput.
-
WSEA - Word set enrichment analysis across multiple corpora for a topic of interest.
-
WSSA - Word set significance analysis for a topic of interest for a given corpus.
-
Directional word2vec - A direction based word2vec model implementation that separates out the vector representations of the left and right contexts of a word. Highlights enrichment of words in either sides of the context.
The WEAVER package is distributed under [GPL - General Public License (>= 2)]
For any questions or comments, please contact [email protected]
I would like to thank Benjamin Schmidt for the incredible wordVectors package which acted as a principal guide in my venture into NLP. This package and many of its functions draw from his package.
I would also like to acknowledge Richard J. So, Lei Sun, Matthew Stephens, Michael Dawson and Allison Garner for helpful discussions and many of the functions in this package developed from the questions and comments they posed in front of me.