A multiple bayes classification solution to word sense disambiguation

Abstract

Word sense disambiguation is a crucial and difficult problem in natural language processing. The problem of word sense disambiguation can be considered as a typical classification problem. Different information is selected to build three classifiers based on Naive Bayes. After forming the confusion matrix to show the ability of each classifier to each sense item as the pre-probability, we compose three classifiers to a multiple classifier system. The result of experiments shows that the multiple classifier system outperforms individual classifier.