Annotating Web Pages for Semantic Web

Adding semantic annotation for Web page is the foundation on constructing semantic Web. Lexical patterns based annotation methods are adopted by most of the semantic annotation systems. The structures and visual features of Web pages imply valuable semantic information, such as relation between entities, but these features are often ignored by current semantic annotation methods. A hybrid pattern called XPattern that represents the background knowledge information and patterns on both lexical and structure of relations to be annotated in Web pages is devised, and an XPattern-based annotation method is proposed in this paper. Experiments show the method can catch more instantiations for both concepts and relations.