Study On Acquiring Non-taxonomic Relations From Wikipedia-like Encyclopedic Web Sites

Ontology, the core concept of Semantic Web, is a descriptive formalization of concepts, relations and instances of them. Ontology Learning is a vital methodology that aims at learning and extracting ontology out from implicit and not-well-structured resources, such as plain text and web pages. This learning method outweighs manual method due to the latter one would be time-consuming and error-prone. By applying ontology learning, taxonomic (sub-concept-of) relations have been largely studied so far but leave non-taxonomic ones not sufficiently considered. Consequently, existing ontologies are mostly restrained to concepts and taxonomic relations among them which have baffled the development of Semantic Web and its applications.As the World Wide Web evolves, there are many resources that are collaboratively edited by ordinary web users. Due to this mechanism, this kind of web sites naturally owns a series of merits, such as knowledge coverage and accuracy. Wikipedia (www.wikipedia.org) is a typical one of them. Its openness and...