A new approach to compute semantic relevance of Chinese words

Words semantic relevance computation which explores semantic relevance/distance between two words is broadly used in many applications, such as word sense disambiguation, information retrieval, structural disambiguation in parsing syntactic, text categorization, etc. We study the relation among sememes and the functions of dynamic roles in concepts and their relevant sememes set. We also propose a new approach to compute words semantic relevance which is based on the relation files provided in HowNet. Experimental results show that this approach can reveal the semantic relevance between words efficiently.