Research on Dual Pattern of Unsupervised and Supervised Word Sense Disambiguation

Abstract

As an important work in the field of natural language processing, word sense disambiguation (WSD) has been a research focus since 1950. The task of WSD is very difficult to solve, and most of modern algorithms fail to reach an ideal level. The processing for WSD is to determine the sense of a polysemous word within a specific context, which involves two steps - determining all the senses for the polysemous word and selecting the appropriate sense among them. In this paper, a dual pattern of WSD based on supervised and unsupervised learning is proposed. Hence, WSD problem can be solved under different circumstances and conditions. Also, an adapted extended Lesk algorithm is established. The experiment results show that the whole quality of unsupervised and supervised WSD is satisfactory