摘要
基于改进的SNN相似度矩阵与谱平分法,提出了一种寻找复杂网络社团结构的算法。首先计算出网络中各节点之间改进的SNN矩阵并将其标准化,求得该矩阵的特征值及特征向量。然后分别选取不同数目的第一非平凡特征向量作为聚类样本,利用FCM聚类算法对节点进行分类,并计算出每次分类结果所对应的模块度Q值。Q的最大值对应的社团结构即为最佳的网络社团结构。一些实验测试了该方法的可行性,通过与其它方法的结果进行比较,可知该算法划分社团的准确率较高。
Based on the improved SNN similarity matrix and spectral bisection method, this paper proposed a new algorithm for detecting the community structure in complex networks. The improved SNN similarity matrix was firstly computed and normalized and its eigenvalues and eigenvectors were obtained subsequently. Then different numbers of the first non-trivial eigenvectors were chosen as clustering samples, FCM algorithm began to work and the corresponding modularity was computed. The best structure of the network was detected by mapping the largest value of modularity. The experiment shows the validity of the presented method. The result obtained here is compared with other popular ones and the conclusion is that the accuracy of the results calculated by this approach is much better than the known ones.
出处
《计算机科学》
CSCD
北大核心
2009年第11期186-188,共3页
Computer Science
基金
国家自然科学基金(10771092)
‘973’项目(2004CB318000)资助
关键词
复杂网络
社团结构
SNN相似度矩阵
谱评分法
FCM算法
Complex networks, Community structure, SNN similarity matrix, Spectral bisection method, FCM algorithm
