应用科学学报 ›› 2017, Vol. 35 ›› Issue (5): 602-611.doi: 10.3969/j.issn.0255-8297.2017.05.006

• 2016中国计算机应用大会遴选论文 • 上一篇    下一篇

基于活跃点的社区跟踪算法

杨绍文, 闫光辉, 李雷, 张海涛   

  1. 兰州交通大学 电子信息学院, 兰州 730070
  • 收稿日期:2016-10-21 修回日期:2016-12-10 出版日期:2017-09-30 发布日期:2017-09-30
  • 通信作者: 闫光辉,教授,研究方向:复杂网络,E-mail:yangh9805@qq.com E-mail:yangh9805@qq.com
  • 基金资助:

    国家自然科学基金(No.61163010);兰州市科技计划项目基金(No.2014-1-171);金川公司预研基金(No.JCYY2013012)资助

Community Tracking Algorithm Based on Active Points

YANG Shao-wen, YAN Guang-hui, LI Lei, ZHANG Hai-tao   

  1. Electronic Information College, Lanzhou Jiaotong University, Lanzhou 730070, China
  • Received:2016-10-21 Revised:2016-12-10 Online:2017-09-30 Published:2017-09-30

摘要:

针对复杂网络社区跟踪中存在忽略演化时域因素以及忽略网络成员演化差异性不足等问题,提出一种社区跟踪方法.对相似函数添加时域信息,并考虑网络演化的平滑性与节点间的差异性,提取网络中的活跃节点进行社区跟踪.实验表明,该算法在DBLP数据集上能比其他社区跟踪算法更好地发现社区演化过程,且找到的社区信息相似度较高.

关键词: 社区跟踪, 活跃节点, 社会网络分析, 社区演化, 时域网络

Abstract:

The research of complex networks is mainly aimed at the complex systems, and it is a general method for dealing with various problems in complex systems. In general, complex network community tracking neglects evolutionary time domain factors and differences in the evolution of network members. This paper proposes a community tracking method that includes time domain information in the similarity function, and extracts active nodes in the network by taking into account smoothness of network evolution and differences between nodes. Experiments show that the proposed algorithm fnds the community evolution process better than those based on DBLP data sets. It can also discover community similarity effectively.

Key words: community evolution, social network analysis, activity point, temporary networks, community track

中图分类号: