A Semantic Relatedness Measure Based on Combined Encyclopedic, Ontological and Collocational Knowledge

TL;DR

This paper introduces a new semantic relatedness measure that combines encyclopedic, ontological, and collocational knowledge sources, achieving state-of-the-art results on the WS-353 benchmark.

Contribution

It presents a novel combined measure integrating Wikipedia, WordNet, and collocation data, improving semantic relatedness accuracy over previous methods.

Findings

Achieved a Spearman rho of 0.79 directly and 0.87 with SVM prediction on WS-353.

Outperformed existing relatedness measures in benchmark tests.

Discussed adaptations and unsuccessful enhancements of ESA.

Abstract

We describe a new semantic relatedness measure combining the Wikipedia-based Explicit Semantic Analysis measure, the WordNet path measure and the mixed collocation index. Our measure achieves the currently highest results on the WS-353 test: a Spearman rho coefficient of 0.79 (vs. 0.75 in (Gabrilovich and Markovitch, 2007)) when applying the measure directly, and a value of 0.87 (vs. 0.78 in (Agirre et al., 2009)) when using the prediction of a polynomial SVM classifier trained on our measure. In the appendix we discuss the adaptation of ESA to 2011 Wikipedia data, as well as various unsuccessful attempts to enhance ESA by filtering at word, sentence, and section level.