Acquisition of Hypernyms and Hyponyms from the WWW
Ratanachai Sombatsrisomboon, Yutaka Matsuo, Mitsuru Ishizuka · 2003
Recently research in automatic ontology construction has become a hot topic, because of the vision that ontology will be the core component to realize the semantic web. This paper presents a method to automatically construct ontology by mining the web. We introduce an algorithm to automatically acquire hypernyms and hyponyms for any given lexical term using search engine and natural language processing techniques. First, query phrase is constructed using the frame “X is a/an Y”. Then corpora sentences is obtained from the result from search engine. Natural language processing techniques is then employed to discover hypernym/hyponym lexical terms. The methodologies proposed here in this paper can be used to automatically augment natural language ontology, such as WordNet, domain specific knowledge.