基于节点属性的社群结构探测算法改进

Improved community structure detection algorithm based on the node's property

对Vincent D.Blondel等提出的B算法的特点及机理进行了分析,讨论了节点属性对社群结构探测的可能影响.进而通过重构初始化网络,控制节点(社群)合并过程两个方面,对B算法进行了改进,获得更优的模块性指标及对应的社群划分.经计算机模拟网络与实际网络的社群结构探测,结果表明所提改进算法有效可用,能在获得较大模块性指标的同时,获得较好的社群划分结果,且拥有更低的运算时间.

This paper analyzes the features and mechanism of the B algorithm proposed by Vincent D. Blondel et al., and discusses the possible impact of nodes＇ property on community structure detection. Then it proposes an improved algorithm for the B algorithm by reconstructing the initial network and controlling node （community） merging process in order to obtain better modularity and the corresponding network partition. The community structure detection experiments on computer simulation networks and actual networks, show that the improved algorithm we proposed is reliable and effective, which achieves a better network partition with a larger modularity and has shorter computation time.