一种移动网络中的重叠社区发现算法.doc

文档介绍：
一种移动网络中的重叠社区发现算法
龚胜芳1,2,孟祥武1,2**
(1. 北京邮电大学智能通信软件与多媒体北京市重点实验室北京 100876;
5
10
2. 北京邮电大学计算机学院北京 100876)
摘要:该文针对移动通信网中的社区发现问题,结合传统社区发现方法的思想提出一种移动
网络中的重叠社区发现算法。利用移动用户行为来构建移动社会化网络模型,给出移动网络
中的节点与社区相似度计算公式,采取限定阈值的 GN 算法对核心图进行初始社区划分;然
后,计算剩余离散节点与各个社区的相似度, 将节点加入到相似度值最大的社区中;重新遍
历所有节点判断是否需要调整,此时加入多个社区的节点表示社区的重叠部分。最后用公开
数据集和模拟数据集的实验结果验证该算法的可行性和有效性。
关键词:移动通信网;社区发现;重叠社区;相似度
中图分类号:TP393
15
An munities discovery algorithm in mobile
networks
GONG Shengfang1,2, MENG Xiangwu1,2
(1. Beijing Key Laboratory of Intelligent munications Software and Multimedia, Beijing
20
25
30
University of Posts and munications, Beijing, 100876;
2. School puter Science, Beijing University of Posts and munications, Beijing,
100876)
Abstract: Based on the mobile users’ behaviors, a new munity discovery
algorithm is proposed to solve munity discovery issues in work,
which bined with the idea of munity discovery methods. Mobile social
network model is constructed by this algorithm, and it gets munities kernel nodes by GN
algorithm which is defined threshold; and then partial munities kernel nodes are
merged to get the current munity, as partial satisfied nodes are added to the local
community, munity structure is discovered ; then, determining all nodes whether need to
be adjusted, then the node added to two or munities is the overlapping part of the
community; finally,this algorithm