Data Mining And Implementation Of Online Communities

Web communities have made a great success and enjoyed an exponential increase on Internet, and formed a huge social network never seen before, during the development of computer and especially the WWW technique. The Web community as a new fashion for managing cyber communities gives a remarkable impact on every member in traditional society, and brings the greater importance to everyone’s routine life and work style. From the classic definition of Web communities, the mission of Web community management system is effectively managing communities, users’behaviors and relationships. The practice of Web community management systems nowadays has no efficient management member relationships effectively. There are heterogeneous Web communities with different topics or network topologies in a Web community management system. When we recommend the potential relationships to community members, we should take account of not only the strength of the relationship between members, but also the different community topic context existed in such a heterogeneous environment. Furthermore, the relationship recommendation candidates need to be quantified for a top k recommendation priority.Based on the node attributes and network topologies in a community member’s social network, we propose a hybrid relationship recommendation mechanism, which is mixed with document topic mining and community member’s reputation evaluation. We analyze the specific features of community member’s relationships in an environment of Web community with topics, and then improve the technique of relationship mining and recommendation. The relationship mining is based on the similarity of community member’s topics. The mining results then are classified by the topics of communities.This paper focused on the web community member relation analysis on a basis of topic mining. Through design and implement the data collection module, member relation analysis module and member relation query module, the experiment results prove its effectiveness and validity.