Using Information Content to Evaluate Semantic Similarity in a Taxonomy

TL;DR

This paper introduces an information content-based measure for semantic similarity in taxonomies, demonstrating superior correlation with human judgments compared to traditional methods.

Contribution

It proposes a novel information content-based similarity measure and empirically shows its improved performance over edge counting in taxonomy-based semantic similarity.

Findings

Correlation of r=0.79 with human judgments

Outperforms traditional edge counting (r=0.66)

Approaches the upper bound of human consistency (r=0.90)

Abstract

This paper presents a new measure of semantic similarity in an IS-A taxonomy, based on the notion of information content. Experimental evaluation suggests that the measure performs encouragingly well (a correlation of r = 0.79 with a benchmark set of human similarity judgments, with an upper bound of r = 0.90 for human subjects performing the same task), and significantly better than the traditional edge counting approach (r = 0.66).