一种发现交叠社团的快速层次化算法

来源期刊:中南大学学报(自然科学版)2010年第5期

论文作者:彭佳扬 杨路明 王建新 李敏 蔡娟

文章页码:1834 - 1840

关键词:复杂网络;社团连通性;社团发现;层次化;交叠

Key words:complex network; community connection; community detecting; hierarchical; overlapping

摘    要:针对大多数层次聚类算法无法识别实际复杂网络中存在的交叠社区等缺陷,提出1种度量社团间连通性的指标,并在此基础上设计1种发现交叠社团的快速层次化算法F-HOC。F-HOC以社团连通性为依据,用凝聚法对k-团进行弱社团检测、递归合并,以达到网络可交叠层次化快速聚类的目的。采用人们普遍接受的基准随机网络作为标准数据对算法进行测试,并应用该算法对足球网络进行分解。研究结果表明:与目前可以发现交叠社团的层次化算法EAGLE相比,对于社团结构明显的复杂网络,F-HOC具有更大的敏感度和更高的运行效率;随着大规模网络数据的不断增加,EAGLE的运行时间呈指数增长,而F-HOC保持线性增长,F-HOC更适用于大规模的复杂网络。

Abstract: Based on the fact that most of the hierarchical clustering algorithm cannot detect overlapping community in complex networks, a new measurement for evaluating the connection of communities was proposed. Based on the proposed measurement, a fast hierarchical clustering algorithm F-HOC was developed to detect overlapping and hierarchical community structure. Maximal cliques containing at least k nodes were generated, and all cliques were merged based on the community connections until they were the weak communities. A node can be in several cliques, and so the community detected by F-HOC can be overlapped. Finally, F-HOC was tested based on the benchmark random network data and the football network. The results show that compared with EAGLE algorithm for detecting overlapping and hierarchical community structure in network, F-HOC can achieve better performances in speed and sensitivity. The operating speed of EAGLE grows exponentially with the increase of network data, while F-HOC maintains linear growth, which makes F-HOC more suitable for large-scale complex networks than EAGLE.

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号