WMD Wrapper Implementation

This is an implementation of Wrapper for WMD (Word Mover's Distance) based on gensim lib in python.

Requirement

• Python3 and supporting packages: Gensim, Numpy, Pandas, NLTK.
• Trained doc2vec or word2vec model

Usage

Run wmd.py with following args:

• Docs directories -l <testdocs dir>;
• Model location -m <model dir>;
• Aggregation threshold -t <threshold>;
• Output file -o <output.csv>.

Output: Clustered results in specified output file.

Time Complexity

$\Theta(n^2)$

Reference

• Matt J. Kusner, Yu Sun, Nicholas I. Kolkin, and Kilian Q. Weinberger. 2015. From word embeddings to document distances.

  To cite using the following BibTeX entry (instead of Google Scholar):

    @inproceedings{kusner2015doc, 
       title={From Word Embeddings To Document Distances}, 
       author={Kusner, M. J. and Sun, Y. and Kolkin, N. I. and Weinberger, K. Q.}, 
       booktitle={ICML}, 
       year={2015}, 
    } 
  

• gensim.

• Weighted graphs using NetworkX