An Integrated Method for Web Resource Categorization

Since Semantic Web is increasing in size and variety of resources, it is difficult for users to find the information that they really need. In this paper, we propose a new approach of integrating four processes for Web resource categorization which can promote the efficiency of information retrieval. The processes can extract both the explicit relations extracted directly from the ontologies in a traditional way and the potential relations inferred from existing ontologies by focusing on some new challenges such as extracting important class names, using WordNet relations and detecting the methods of describing the Web resources. Our proposed method can achieve a notable improvement in categorizing the valuable Web resources based on incomplete ontologies for a Semantic Web search system.