Although community discovery based on social network has been studied extensively in the Web hyperlink environment, limited research has been done in the case of Web documents. The co-occurrence of words and entities in sentences and documents usually implies some connections among them. Studying such connections may reveal important relationships. In this paper, we investigate the co-occurrences of named entities in Web pages and blogs, and mine communities among those entities. We show that identifying communities in such an environment can be transformed into a graph clustering problem. A hierarchical clustering algorithm is then proposed, which exploits triangle structures within the graph and the mutual information between vertices. Our empirical study shows that the proposed algorithm is promising in discovering communities from Web documents.

PDF

Luo, X., Kenyonm R. V., Guan, Y., Pervasive Web Community Structure Summarization: A Machine Learning Approach, In Proceedings of the 2006 International Conference on Machine Learning; Models, Technologies & Applications (MLMTA ’06), Las Vegas, NV, June 29th, 2006.