A WORD SENSE DISAMBIGUATION BASED ON X-BAR

A word sense disambiguation based on X bar is introduced. The approach compares the word’s X bar with the usage in the dictionary from two aspects:the structure and the words in the X bar. The words’ similarity is evaluated using WordNet. The approach has the virtue of the demand for a small training set and it can avoid the problems in traditional methods based on word co occurrence, such as the demand for a large corpus, sparse data, etc .