基于桥系数的分裂社区检测算法研究
发布时间：2018-11-27 07:36
【摘要】：研究社区结构有助于揭示网络结构和功能之间的关系,而社区检测是社区结构研究的基础和核心。该文定义了一种聚集度桥系数,将其应用到社区检测中,设计出一种分裂社区检测方法,包括分裂和合并两个算法。分裂算法使用桥系数识别社区间边,通过迭代删除社区间边分解网络,从而发现网络中的社区结构;合并算法根据社区连接强度合并社区,可以揭示社区结构中的分层嵌套的现象。在六个社会网络数据集上的实验表明,本文算法可以有效的将网络分裂为有意义的社区,并且准确性接近或超过经典的社区检测算法。
[Abstract]:The study of community structure is helpful to reveal the relationship between network structure and function, and community detection is the foundation and core of community structure research. In this paper, a clustering bridge coefficient is defined and applied to community detection, and a split community detection method is designed, which includes two algorithms: split and merge. The split algorithm uses the bridge coefficient to identify the inter-community edges and iteratively deletes the inter-community edge decomposition network so as to find the community structure in the network. The merging algorithm can reveal the phenomenon of stratification and nesting in the community structure according to the intensity of community connection. Experiments on six social network datasets show that the proposed algorithm can effectively split the network into meaningful communities, and the accuracy is close to or higher than the classical community detection algorithms.
【作者单位】： 山西大学计算机与信息技术学院;
【基金】：国家自然科学基金(61175067,61272095,61432011,61573231) 山西省科技基础条件平台计划项目(2015091001-0102) 山西省回国留学人员科研项目(2013-014)
【分类号】：O157.5
本文编号：2359890
[Abstract]:The study of community structure is helpful to reveal the relationship between network structure and function, and community detection is the foundation and core of community structure research. In this paper, a clustering bridge coefficient is defined and applied to community detection, and a split community detection method is designed, which includes two algorithms: split and merge. The split algorithm uses the bridge coefficient to identify the inter-community edges and iteratively deletes the inter-community edge decomposition network so as to find the community structure in the network. The merging algorithm can reveal the phenomenon of stratification and nesting in the community structure according to the intensity of community connection. Experiments on six social network datasets show that the proposed algorithm can effectively split the network into meaningful communities, and the accuracy is close to or higher than the classical community detection algorithms.
【作者单位】： 山西大学计算机与信息技术学院;
【基金】：国家自然科学基金(61175067,61272095,61432011,61573231) 山西省科技基础条件平台计划项目(2015091001-0102) 山西省回国留学人员科研项目(2013-014)
【分类号】：O157.5
【相似文献】
相关期刊论文 前6条
1 庞明勇,卢章平;局部数据集与噪声数据曲线的平滑过滤[J];矿山测量;2001年04期
2 吴国清;莫则尧;陈虹;;一种基于信息测度的科学数据集序列约减方法[J];计算物理;2009年06期
3 柳金甫;怎样处理异常数据[J];中国统计;1995年07期
4 易昆南;袁中萸;;对模拟纵向数据集缺失值处理的几种方法比较[J];湖南工业大学学报;2008年02期
5 朱梅红;;一种加权的MCLP分类模型及其在不平衡数据集上的应用[J];统计与决策;2011年21期
6 崔玉泉,马建华,李振波;对一类数据集识别问题的探究[J];山东大学学报(理学版);2004年05期
相关硕士学位论文 前2条
1 张健;剪枝和网格采样相结合的非平衡数据集分类方法[D];安徽大学;2012年
2 姜君;CIECAM02的若干问题及解决办法[D];辽宁科技大学;2015年
，本文编号：2359890
本文链接：https://www.wllwen.com/kejilunwen/yysx/2359890.html
