基于相似度的三元社团合并算法  

Ternary community merging algorithm based on similarity degree

作　　者：吴磊[1] 谢刚[1] 杨云云[1] 

机构地区：[1]太原理工大学信息工程学院,太原030024

出　　处：《计算机应用研究》2016年第11期3270-3273,3278,共5页Application Research of Computers

基　　金：国家留学回国人员科技活动择优资助项目([2012]258号);山西省留学回国人员科技活动择优资助项目([2012]316号)

摘　　要：针对使用相似度测量进行社团划分时可能出现的判断冲突问题,提出了一种基于相似度的三元社团合并算法。首先计算网络中所有节点相似度,并构建相似度矩阵和阈值矩阵。通过对相似度阈值的选取,筛选网络中不同的三元社团,并将其作为社团合并的基本元素,通过社团相似度将其合并。然后将剩余节点和孤立三元社团分别按照节点从属度和三元社团从属度划分到相应社团。三元社团的构建更加清晰地凸显了社团结构。最后通过在人工合成网络和真实世界网络上进行实验测试,结果表明算法可以准确高效地将网络中的节点划分到相应的社团。For the possible judgment conflict problem in community detection algorithms based on similarity, this paper pro- posed a ternary community merging algorithm based on similarity. Firstly, it calculated the similarity of all nodes in the net- work and constructed the similarity matrix and threshold matrix. Through the selection of the similarity threshold, it selected the different ternary communities in the network as the basic element and merged it through community similarity. Then it di- vided the remaining nodes and isolated ternary community into the corresponding community respectively in accordance with the degree of membership of node and the degree of membership of the ternary community. The construction of the ternary com- munity displayed the community structure more clearly. In the end, the results show that the nodes in the network can be di- vided into the corresponding groups accurately and efficiently by using the method in the artificial synthetic network and real world network.

关 键 词：复杂网络 社团发现 相似度 从属度 三元社团 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]