A word sense disambiguation system based on bayesian model

Research on word sense disambiguation (WSD) is of great importance in natural language processing fields. In this paper, a novel word sense disambiguation system is designed in which bayesian theory is applied to determine correct sense of an ambiguous word. Morphology knowledge in word unit is mined to guide WSD process. Neighboring morphology knowledge of an ambiguous word is used as feature for constructing WSD classifier. Word segmentation tool is integrated into this system and browser/server (B/S) framework is adopted. Experimental results show that the performance of WSD system is good.