A matrix approach for hierarchical web page clustering based in hyperlinks

This paper proposes a matrix approach for hierarchical web page clustering with two algorithms using hyperlink information among pages. One clustering algorithm clusters web pages without considering cluster overlapping. Another one takes cluster overlapping into account. These algorithms take advantage of intrinsic relationships among the pages, and are independent of the order in which the pages are presented to the algorithms. Furthermore, the proposed algorithms do not require a predefined similarity threshold for clustering. They are easy to be implemented for web applications. The primary evaluations show the effectiveness of the proposed algorithms, as well as a promising application.