Abstract
Word sense disambiguation (WSD) is a task in natural language processing, which asks to identify the appropriate sense of a word according to a particular context. Several approaches were investigated to tackle the WSD problem, including genetic algorithms. In this paper, we propose a new genetic algorithm, called GAWSD, that benefits from part-of-speech tagging, domain knowledge, and gloss enrichment to find a sense to a target word. The performance of the algorithm was evaluated on fine-grained and coarse-grained standard corpora. The results show that GAWSD outperformed the best known algorithms on the fine-grained corpus. This result sets GAWSD as a competitive algorithm for WSD.