摘要
社会网络和复杂网络上的社区识别已经成为当前研究的热点和前沿课题.针对目前社区识别方法不能兼具较低时间复杂度、无须专家知识或先验知识和允许存在重叠节点的不足,提出了基于拓扑势理论的重叠社区识别方法.通过提出的重叠节点社区归属不确定性测度,该方法同时实现了社区间结构洞的识别.实验验证了该方法的有效性.另外,文章在理论证明的基础上提出了影响因子优化算法;论证了结构洞理论视角下网络的脆弱性.
Community identification has been a hot spot and a cutting-edge topic among researchers. Since none of the present community identification methods simultaneously meets the requirements, such as lower time complexity,independence of ex- pertise or experiences, allowance for overlapping nodes and so on, an overlapping community identification method is proposed based on topological potential theory. This method can also identify the structural holes in communities at the same time by the presented uncertainty measure of the community identity of the overlapping nodes, and its effectiveness is verified by experiments. In addition, an influence factor optimization algorithm is proposed and network fragility is discussed and prooved from the perspective of structural hole theory.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2014年第1期62-69,共8页
Acta Electronica Sinica
基金
国家自然科学基金(No.61073041
No.61073043)
黑龙江省自然科学基金(No.F200901
No.F200917)
黑龙江省教育厅科学技术研究基金(No.12531529)
哈尔滨市优秀学科带头人基金(No.2010RFXXG002
No.2011RFXXG015)
高等学校博士学科点专项科研基金(No.20112304110011)
关键词
网络
重叠社区
结构洞
识别
拓扑势
影响因子
不确定性测度
脆弱性
network
overlapping community
structural holes
identification
topological potential
influence factor
uncertainty measure
fragility