摘要： 针对复杂网络中难以发现小社区的问题，在CNM算法的基础上，提出一种利用局部信息进行社区挖掘的方法。定义节点的强度及节点对社区的贡献，改进模块度使该方法能适用于带权网络。利用社区局部信息得到小社区集合，将小社区集合作为CNM算法的输入，计算小社区间的模块度增量，凝聚模块度增量小的小社区，并得到最终结果。实验结果表明，该方法具有较高的社区模块度和算法执行 效率。

关键词: 复杂网络, 社区发现, 聚类, 加权模块度, 图分割

Abstract: It is difficult to find small community in complex network. Aiming at this problem, this paper presents a community detection method using local information based on CNM algorithm. Several measures are adopted such as the definition of both the vertices strength and the vertex’s contribution to community, the improvement of modularity of community structure making the method useful for weighted network, and the calculation of small community set with local information of community. It uses the small community to set as the input of CNM algorithm, the method gets the result by computing the increment of modularity among small communities and merging the small communities with minimum increment. Experimental results show that the modularity of community structure found by the method and the algorithm efficiency are higher than the state-of-the-art algorithms.

Key words: complex network, community detection, clustering, weighted modularity, graph partition

中图分类号: 

• TP18