Learning a Robust Word Sense Disambiguation Model using Hypernyms in Definition Sentences

This paper proposes a method to improve the robustness of a word sense disambiguation (WSD) system for Japanese. Two WSD classifiers are trained from a word sense-tagged corpus: one is a classifier obtained by supervised learning, the other is a classifier using hypernyms extracted from definition sentences in a dictionary. The former will be suitable for the disambiguation of high frequency words, while the latter is appropriate for low frequency words. A robust WSD system will be constructed by combining these two classifiers. In our experiments, the F-measure and applicability of our proposed method were 3.4% and 10% greater, respectively, compared with a single classifier obtained by supervised learning.