Research of Word Sense Disambiguation Based on Soft Pattern

Word sense disambiguation (WSD) is always an important and difficult problem that requires to be solved in Nature Language Processing. This paper presents a new WSD method which is based on soft pattern matching. The method can learn the soft patterns from the sense of the ambiguous word and its context, to construct a soft pattern - based database. At last the sense of the ambiguous word is labeled by choosing the sense with the maximum matching degree between the ambiguous word context and the soft pattern. The experiment result shows that the method has high precision.