Betweenness research in telecom social network

The research on static measure betweenness of telecom social network has some great challenges which include the research of the distribution law of betweenness and the computing of betweenness in a giant network. With massive telecommunication data, we have done some research about betweenness of telecom social networks, which includes three parts. First part is about the natural properties of betweenness. Our research demonstrates that the distribution of betweenness conforms to the Power-Law. Some results about the relation between betweenness and other static measures are presented as well. Second part is about how to reduce the computation time of betweenness with the help of high performance computing systems. Our results show that the computation time is almost linear with the number of CPUs with a speedup ratio 53.93. Third part is about the approximate algorithm for the betweenness computation. We present the precisions of betweenness for each node in diverse situations with approximate algorithm, and rind that the number of hops in the shortest path is positively proportion to the computation precision of betweenness.