Chinese Verb Sense Disambiguation Using AdaBoosting

This paper uses the adaptive boosting (AdaBoosting) algorithm to the task of word sense disambiguation (WSD) for Chinese verbs. The AdaBoosting algorithm is a kind of ensemble learning method used for classification. We have implemented the classifier using a feature set combining collocation features, syntactic features and semantic features. We test the model on eight polysemous verbs in Chinese and show that for most of the verbs, performance of WSD task can be improved by adding more features. The senses of some highly polysemous verbs are hard to distinguish for there are not too many samples for each sense.