Improving Word Similarity by Augmenting PMI with Estimates of Word Polysemy

Authors: Lushan Han, Tim Finin, Paul McNamee, Anupam Joshi, and Yelena Yesha

Journal: IEEE Transactions on Knowledge and Data Engineering (TKDE)

Date: February 01, 2012

Abstract: Pointwise mutual information (PMI) is a widely used word similarity measure, but it lacks a clear explanation of how it works. We explore how PMI differs from distributional similarity, and we introduce a novel metric, PMImax, that augments PMI with information about a word's number of senses. The coefficients of PMImax are determined empirically by maximizing a utility function based on the performance of automatic thesaurus generation. We show that it outperforms traditional PMI in the application of automatic thesaurus generation and in two word similarity benchmark tasks: human similarity ratings and TOEFL synonym questions. PMImax achieves a correlation coefficient comparable to the best knowledge-based approaches on the Miller-Charles similarity rating dataset.

Type: Article

Publisher: IEEE Computer Society

Note: (to appear)

Tags: semantics, semantic similarity, pointwise mutual information, word similarity, ontology, synonyms

Google Scholar: search

Number of downloads: 150

Available for download as

size: 442906 bytes

Related Projects:

Active Project

 Graph of Relations.

Bookmark at: Digg | Del.icio.us | Connotea | CiteULike