An Approach For Discovering Web Community Based On Weighted Betweenness

Web is a typical complex network. There exists a lot of communities while Web grows. These communities are very important information in Web organization. Communities can provide valuable and credible information in time for users, Web community can make one knows the knowledge information in the Web and it's organization structure status. Web community represent the social activities in the web. Finding Web communities is a significant methodology to improve query-time efficiency and classify Web pages exactly.This thesis classifies the popular community identification techniques presently, and implements these methods. We find these methods appear topic drift Phenomenon. Based on web's complex network characteristic. We provide an improved GN algorithm to find web communities on special topics. The Levenshtein distance is employed to metric the similarity between a Web page and the topic and every two Web pages. Then use GN algorithm to divide the network graph. Improved GN algorithm's complex is drop clearly. We provide community density and community average similarity to guide us to choose a good community. The result of the experiment shows this approach can find high quality communities on specific topic.