An Exhaustive and Edge-Removal Algorithm to Find Cores in Implicit Communities

Web community is intensely studied in web resource discovery. Many literatures use core as the signature of a community. A core is a complete bipartite graphs, denoted as Ci,j. But discovery of all possible Ci,j in the web is a challenging job. This work has been investigated by trawling [1][2]. Trawling employs repeated elimination/generation procedure until the graph is pruned to a satisfied state and then enumerate all possible Ci,j. We proposed a new method that uses exhaustive and edge removal method. Our algorithm avoids scanning dataset many times. Also, we improve crawling method by only recording potential fans to save disk space. The experiment result show that the new algorithm works properly and many new Ci,j can be found by our method.