一种全局与局部结合的两步式重叠社区挖掘算法

A two-step overlap community mining algorithm based on global and local methods

基于全局划分和局部凝聚原理,改进得到一种两步式挖掘算法,该算法以寻找最优模块性Q值为基准,最终挖掘出重叠社区.对两个经典真实世界网络的Zachary's Karate俱乐部数据和海豚网络数据进行了实验测试,实验表明该算法能够有效地划分出重叠社区.

Based on global partition and local agglomeration, this paper presents an algorithm to be carried out in two steps. The algorithm is to find the optimal modularity Q value as a benchmark. Finally,overlap communities can be obtained. Two typical real world data sets, Zachary ＇s Karate club data and dolphin network data, were tested by experiments. The result showed that the algorithm could help to mine out overlapping community.