可多边并行移出的社团发现方法  

Community Detection Method with Multi-edge Simultaneous Removal

作　　者：熊中敏[1,2] 黄冬梅[2] 

机构地区：[1]复旦大学计算机与信息技术系,上海200433 [2]上海海洋大学信息学院,上海201306

出　　处：《计算机工程》2009年第12期29-31,共3页Computer Engineering

基　　金：国家科技部海洋公益性行业科研专项经费基金资助项目(200805016);国家"973"计划基金资助项目(2005CB321905);国家自然科学基金资助项目(60303008);上海海洋大学博士启动基金资助项目

摘　　要：针对GN算法计算效率低下的缺陷,提出一个基于边的中介值测度的发现网络潜在社团结构的新算法。该算法在完成所有边的中介值计算后,利用成分的独立性,采用并行移出各个成分中具有最大中介值的边的方法。通过理论分析,在作为实验测试平台的实际的数据集上进行实验验证,结果表明该算法是快速、有效的。To address the slow speed of GN algorithm, a new algorithm based on betweenness scores of edges is presented for detecting the underlying community structure in networks. Employing component independency, this algorithm presents a new method through which all edges with the highest betweenness score in respect of each component is simultaneously removed when all betweenness scores are computed. It is proved that this algorithm is fast and effective through theoretical analysis and experiments with several real data sets which are acted as test beds.

关 键 词：社团发现 社会网络 社团结构 图挖掘 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]