Weighted Graph Clustering for Community Detection of Large Social Networks

This study mainly focuses on the methodology of weighted graph clustering with the purpose of community detection for large scale networks such as the users’ relationship on Internet social networks. Most of the networks in the real world are weighted networks, so we proposed a graph clustering algorithm based on the concept of density and attractiveness for weighted networks, including node weight and edge weight. With deep analysis on the Sina micro-blog user network and Renren social network, we defined the user's core degree as node weight and users’ attractiveness as edge weight, experiments of community detection were done with the algorithm, the results verify the effectiveness and reliability of the algorithm. The algorithm is designed to make some breakthrough on the time complexity of Internet community detection algorithm, because the research is for large social networks. And the another advantage is that the method does not require to specify the number of clusters.