Finding Communities Using User Preference in Web Structure Mining

摘要

Web structure mining is the method based on the graph structure of hyperlinks, and it does not use the information of web contents. HITS algorithm and PageRank algorithm are popular methods for web structure mining. In this study, we deal with the finding algorithm of web communities in web structure mining. This algorithm receives some URLs of user's known web pages, and proposes to the user the candidates of pages in the web community by using the structure of bipartite graph. We investigate the effect of introduction of user preference to each known pages, and discuss the way to improve the finding algorithm of web communities.