General-Purpose Ontology Enrichment from the WWW

Abstract

Ontology enrichment is required when the knowledge captured by the ontology is out of date or unable to capture the specified user requirements in a specific domain. In this paper we present an automatic statistical/semantic framework for enriching general-purpose ontologies from the World Wide Web (WWW). Using the massive amount of information encoded in texts on the web as a corpus, missing background knowledge such as concepts, instances and relations can be discovered and exploited to enrich general-purpose ontologies. The benefits of our approach are: (i) enabling ontology enrichment with missing background knowledge, and thus, enabling the reuse of such knowledge in future. (ii) saving time and effort required to manually enrich and update the ontologies. Experimental results indicate that the techniques used to enrich ontologies are both effective and efficient.