Structural analysis and grouping of Web pages

In order to easily cope with scattered information on the Web, we propose a method of grouping Web pages on a site. This method allows us to decompose the whole structure of a site by considering semantically related pages as one virtual document. In this paper, we describe the proposed grouping method of Web pages based on the link structure between pages without using similarity between documents which is utilized by traditional text categorization. We consider a Web page set as a Web graph and try to extract strongly connected components as groups. Next, because groups comprise hierarchy structure, we divide a strongly connected component to extract the hierarchy structure.