二谱分聚类的复杂网络社团检测方法

近年来,复杂网络中的社团发现越来越受到研究人员的关注并且许多方法被提了出来。为有效地检测复杂网络中的社团结构,优化了评估与发现社团的模块密度函数(即D值)。通过模块密度的优化进程,证明了模块密度函数能写成模块密度矩阵迹的最大化表达形式。利用模块密度矩阵的谱分分解,提出了一种新的二谱分的聚类检测复杂网络社团方法。在LFR标准人工模型网络中验证了二谱分方法的有效性。实验结果显示这种新的方法在发现复杂网络社团上有较高的准确性。

Detecting Community Structures in Complex Networks:A Bisection Spectral Clustering Approach

In recent years, the problem of community structure detection has attracted more and more attention and many approaches have been proposed. To detect community structures in complex networks, the modularity density function (D value) is optimized, by optimizing process, the modularity density function can be expressed as a trace maximization form about the modularity density matrix. By spectral optimization of modularity density matrix, a bisection spectral clustering approach is proposed to detect communities in complex networks. The algorithm is validated in the LFR benchmark networks. Experimental results show the significance of the proposed approach.