Discovery of web communities from positive and negative examples

Abstract

Several attempts have been made for Web structure mining whose goals are to discover Web communities or to rank important pages based on the graph structure of hyperlinks. Discovery of Web communities, groups of related Web pages sharing common interests, is important for assisting users' information retrieval from the Web. There are several different granularities of overlapping Web communities, and this makes the identification of objective boundaries of Web communities difficult. This paper proposes a method for discovering Web communities from given positive and negative examples. Since the boundary of a Web community is hard to define only from positive examples, negative examples are used for limiting its boundary from outer side of the Web community. Experimental results are shown and the effectiveness of our new method is discussed. © Springer-Verlag Berlin Heidelberg 2003.