Random Walks for Knowledge-Based Word Sense Disambiguation

Abstract

Word Sense Disambiguation (WSD) systems automatically
choose the intended meaning of a word in context. In this article we
present an algorithm based on random walks over large Lexical
Knowledge Bases (LKB) for efficient and state-of-the-art WSD. Using
WordNet we show that the algorithm performs better than previous
unsupervised approaches using similar knowledge on a variety of
datasets, including two specific domains and an additional
language. We also describe a method to apply our algorithm to
domain-specific words which outperforms supervised WSD
systems. Finally, we include a detailed analysis of the factors that
affect the algorithm. The algorithm and the LKBs used are publicly
available, and the results easily reproducible.