DIT: Summarisation and Semantic Expansion in Evaluating Semantic Similarity

Abstract

This paper describes an approach to implementing a tool for evaluating semantic similarity. We investigated the potential benefits of (1) using text summarisation to narrow down the comparison to the most important concepts in both texts, and (2) leveraging WordNet information to increase usefulness of cosine comparisons of short texts. In our experiments, text summarisation using a graph-based algorithm did not prove to be helpful. Semantic and lexical expansion based upon word relationships defined in WordNet increased the agreement of cosine similarity values with human similarity judgements.