基于中心节点和局部优化的复杂网络社团划分方法

A Community Partition Method of Complex Network Based on Central Node and Local Optimization

复杂网络的社团划分是复杂网络研究的重要分支,常用的社团挖掘算法主要分为全局社团挖掘算法和局部社团挖掘算法。针对全局社团挖掘算法难以适应规模和动态性愈来愈大的复杂网络的问题,提出一种基于中心节点和局部优化的复杂网络社团划分方法。首先综合判断节点的度中心性、介数中心性、接近中心性和相似度来确定社团的中心节点集,然后根据局部模块优度将节点划归至合适社团,再对孤立节点和重叠社团节点进行处理,最终实现复杂网络的社团划分。通过实验并与典型局部社团挖掘算法进行对比,所提方法可以实现社团的有效和准确划分。

Community partition of complex networks is an important branch of complex networksresearch, and common community mining algorithms are mainly divided into global community miningalgorithm and local community mining algorithm. Aiming at the problem that the global communitymining algorithm is difficult to adapt to complex networks with increasing scale and dynamics, this paperproposes a community partition method for complex networks based on central node and localoptimization. At first, the center node set of a community is determined by comprehensively judging thedegree centrality, betweenness centrality, proximity centrality and similarity of nodes, and then nodes areclassified into appropriate communities according to the local module goodness, and then isolated nodesand overlapping community nodes are processed to finally realize community partition of complexnetworks. Through experiments and comparison with typical local community mining algorithms, theproposed method can effectively and accurately divide communities.