Domain-Specific Ontology Construction from Hierarchy Web Documents

Abstract

Ontology, the base of Semantic Web, plays a vital role in knowledge representation and knowledge reasoning. There are many tools providing management interfaces to create, query and edit knowledge of ontology. However, most of them are still arduous time-consuming manual work. Therefore, how to automatically construct ontology has attracted many researchers' attention. With the development of WEB, there is influent information we can take advantage of. According to the structure of Web, An automatically construction method for domain-specific ontology is proposed. Firstly, some special web sites which have relatively structure or semi-structure are selected. Then, web documents are clawed with Jsoap. Secondly, all contained knowledge is extracted and organized together to form the domain-specific ontology. Finally, the Jena platform is employed to create, delete, read, write ontology model in the form of RDF and query by SPARQL. The experimental results in “Entertainment” and “Sport” domain show concept hierarchy structures are reasonable with the overall precision of domain-specific ontology being up to 97%.