J4 ›› 2013, Vol. 26 ›› Issue (5): 61-68.doi: 10.3976/j.issn.1002-4026.2013.05.012

• 论文 • 上一篇    下一篇

基于标签传播的稳定重叠社区挖掘算法研究

王庚,宋传超,盛玉晓,王童童,李盛恩   

  1. 山东建筑大学计算机科学与技术学院,山东 济南 250101
  • 收稿日期:2013-03-21 出版日期:2013-10-20 发布日期:2013-10-20
  • 作者简介:王庚(1987-),男,硕士研究生,研究方向为数据库、社会网络。

Label propagation based detecting algorithm of stable overlapping communities

 WANG Geng, SONG Chuan-Chao, SHENG Yu-Xiao, WANG Tong-Tong, LI Sheng-En   

  1. School of Computer Science and Technology, Shandong Jianzhu University, Jinan 250101, China
  • Received:2013-03-21 Online:2013-10-20 Published:2013-10-20

摘要:

        为解决重叠社区挖掘算法中效率与稳定性不能兼顾的情况,本文提出了一种基于标签传播的稳定重叠社区挖掘算法(SOCLP)。该算法首先利用两种标签选择策略解决了基于标签传播算法中的稳定性问题,然后又通过引入平衡因子对算法的稳定性进行了有效的控制。实验结果证明,本文所提出的SOCLP算法可以同时关注社区挖掘的质量、效率和稳定性。

关键词: 社会网络, 社区挖掘, 重叠社区, 标签传播

Abstract:

       We present a label propagation based detecting algorithm of stable overlapping community to consider both efficiency and stability in detecting algorithm of overlapping community. The algorithm initially employs two label selection strategies to solve stability problem in label propagation based algorithm. It then effectively controls its stability by a balance factor. Experimental results prove that the algorithm can simultanously focus on quality, efficiency and stability of a mining community.

Key words: social networks, community mining, overlapping community, label propagation

中图分类号: 

  • TP311.13