Word Sense Disambiguation with THESSOM

Abstract

Word sense disambiguation automatically determines the appropriate senses of a word in context. We have previously shown that self-organized document maps have properties similar to a large-scale semantic structure that is useful for word sense disambiguation. In this article we formalize THESSOM, which is an algorithm for word sense disambiguation using selforganized document maps created with WEBSOM. The algorithm is tested on the SENSEVAL-2 benchmark data and shown to be on a par with the top three contenders of the SENSEVAL-2 competition. We also show that the performance of the algorithm improves when using more advanced linguistic features for creating the WEBSOM maps.