基于弱派系的多层社会网络重叠社团发现算法

北京信息科技大学信息与通信工程学院, 北京 100101

计算机网络; 多层社会网络; 弱派系; 重叠社团; 社团发现; 复杂网络

Overlapping community detection algorithm based on weak clique in multi-layer social networks
ZHANG Yuexia, YANG Ruiqi, and KANG Jin

School of Information and Communication Engineering, Beijing Information Science & Technology University, Beijing 100101, P.R.China

computer network; multi-layer social network; weak clique; overlapping community; community detection; complex network

DOI: 10.3724/SP.J.1249.2018.04413

备注

针对现有社团发现算法中多层社会网络的重叠社团发现算法较少,且较难检测小型多层网络中社团的问题,提出一种基于弱派系的多层社会网络重叠社团发现算法.算法通过检测与合并网络中的弱派系得到社团发现结果,弱派系的构建综合考虑了节点度和节点邻居间的连接,得到更细粒度的社团结构,并同时适用于无向与有向网络.真实网络的实验结果表明,该算法可有效检测小型多层社会网络中的重叠社团,优于现有的基于局部社团的社团发现算法(local community based community detection algorithm,LC-CDA算法).

There are few overlapping community detection algorithms in the existing community detection algorithms and it is difficult to detect communities in small multi-layer social networks. In order to solve the above-mentioned problems, we propose an overlapping community detection algorithm based on the weak clique in multi-layer social works. The proposed algorithm detects the communities by detecting and merging the weak cliques in the network. The construction of weak clique considers the node degree and the number of links between the neighbor nodes. This method can obtain the finer-grained community structures and is suitable for both undirected and directed networks. The experimental results on real world networks show that this method can detect the overlapping community in the small multi-layer social networks effectively and is superior to the local community based community detection algorithm(LC-CDA).

·