An Improved Community Partition Algorithm Integrating Mutual Information

The research of community detection can help us analyze kinds of problems in social network, in which the research of community structure is very important. This paper proposes an improved algorithm: An Improved BGLL Integrating Mutual Information (BGLLi), which stabilizes modularity, meanwhile fuses the index of mutual information, so that we can find the optimal threshold of modularity and reduce the running time of community partition effectively. The dataset adopts the Twitter dataset and the public Arxiv dataset, and the related experimental results have verified the effectiveness of the algorithm