Learning Taxonomical Relations from Domain Texts Using WordNet and Word Sense Disambiguation

Learning taxonomical relations from domain texts is an important task for ontology learning from texts. We observe that rich information on taxonomical relations is available in the lexical knowledge base WordNet. However, in order to exploit the taxonomical relations in WordNet we need to tackle the difficult problem of word sense disambiguation. In this paper, we present a weighted word sense disambiguation method and show its application for learninng taxonomical relations from domain texts. The experimental results indicate that using Word Net and our word sense disambiguation method achieves good accuracy and coverage for the learning task.