采用模糊层次聚类的社会网络重叠社区检测算法

发布时间：2018-01-23 20:20

本文关键词： 社会网络相似度模糊层次聚类重叠社区检测　出处：《西安交通大学学报》2015年02期 　论文类型：期刊论文

【摘要】：为了能够发现社会网络中的重叠社区以及解决重叠社区之间关系的模糊性和层次性,提出了一种基于模糊层次聚类的重叠社区检测算法(CDHC)。算法中引入了距离加权因子来计算社区间的相似度,通过模糊层次聚类来合并相似度高的社区;针对合并生成的原始社区计算社区中节点的隶属度,再将隶属度小于阈值的节点从社区中移除,从而形成最终的网络重叠社区结构。该算法不仅可以发现重叠的社区结构,还可以处理孤立节点。在Lancichinetti基准网络和真实网络上将CDHC算法与具有代表性的重叠社区发现算法CMP和LFM进行了比较,结果表明:影响社区检测精度的主要因素是社区间的混合程度,而网络规模和网络中社区的规模的影响并不显著;CDHC算法在小社区网络上的社区检测精度优于LFM,在大社区网络上的社区检测精度优于CMP。CDHC算法在保持社区检测质量的同时,还具有较好的稳定性,是一种有效的社会网络重叠社区检测算法。
[Abstract]:In order to find overlapping communities in social networks and resolve the fuzzy and hierarchical relationships between overlapping communities. An overlapping community detection algorithm based on fuzzy hierarchical clustering is proposed. The distance weighting factor is introduced to calculate the similarity between communities. Fuzzy hierarchical clustering is used to merge communities with high similarity; According to the membership degree of the nodes in the original community computing community generated by the merger, the nodes whose membership degree is less than the threshold value are removed from the community. The algorithm can not only find the overlapping community structure. The CDHC algorithm is compared with the typical overlapping community discovery algorithm CMP and LFM on the Lancichinetti benchmark network and real network. The results show that the main factor affecting the accuracy of community detection is the degree of mixing between communities, but the influence of network scale and community size in the network is not significant. The community detection accuracy of CDHC algorithm is better than that of LFM algorithm on small community network, and that of community detection accuracy on large community network is better than that of CMP.CDHC algorithm while maintaining the quality of community detection. It also has good stability and is an effective social network overlapping community detection algorithm.
【作者单位】：西安交通大学电子与信息工程学院;西安交通大学陕西省计算机网络重点实验室;
【基金】：国家自然科学基金资助项目(61172090) 陕西省科技统筹创新工程资助项目(2013SZS16-Z01/P01/K01) 中央高校基本科研业务费专项资金资助项目(XJJ2014049,XKJC2014008) 陕西省自然科学基金资助项目(2012K06-30,2014JQ8322)
【分类号】：TP393.0
【正文快照】： 社会网络是指社会个体成员之间因为互动而形成的相对稳定的关系体系。社会网络通常可抽象为节点以及节点间的边组成的图,图中的节点代表个体,节点间的边表示个体之间的联系。社会网络中蕴含着一些潜在的社区结构,同一社区内的节点具有较高的相似性,而不同社区间的节点相似性较

【参考文献】