基于链路标签传播的重叠社团发现算法

摘要

为挖掘复杂网络中的重叠社团结构，在标签传播算法的基础上，从链路的角度出发，提出一种新的标签更新策略。考虑不同邻居对链路标签贡献值的不同特点，使其更新过程更加快速，克服当前基于节点的标签传播算法需要指定节点所属社团个数的缺陷，在无需人工干预的情况下得到稳定的重叠社团结构。将该算法分别应用于人工网络和真实的社会网络中，实验结果表明，该算法能够快速、有效地挖掘稳定的重叠社团结构。%To identify the overlapping community structure in the complex network ,a novel label propagation method based on the link in the network was presented ,which was compared with the traditional node-based algorithm .This algorithm considered that the neighbors of a link had different dedications when updating the labels of a link .The stable overlapping community wi-thout extra information was gotten by using the new label propagation method .The experimental results demonstrate that the proposed algorithm can detect the overlapping community structure effectively in both artificial networks and real social networks .