Computer Engineering and Applications ›› 2010, Vol. 46 ›› Issue (9): 51-54.DOI: 10.3778/j.issn.1002-8331.2010.09.016

• 研究、探讨 • Previous Articles     Next Articles

Detecting community algorithm based on signal process and hierarchical clustering

HUANG Hao-ying,MA Ying-hong   

  1. School of Management & Economy,Shandong Normal University,Jinan 250014,China
  • Received:2009-10-12 Revised:2010-01-08 Online:2010-03-21 Published:2010-03-21
  • Contact: HUANG Hao-ying



  1. 山东师范大学 管理与经济学院,济南 250014
  • 通讯作者: 黄浩英

Abstract: Community is one of important characters in social networks and community detecting is also a fashionable statement recently.In this paper,based on signaling process on complex networks,influence vectors of each node are got,topological structure of each node is translated into geometrical relationships of vectors in algebra spaces,and by the aid of hierarchical clustering modularity method,communities are detected effectively.With data simulations on the Zachary Karate Club network,College Football network and Dolphin social network,it shows that the proposed algorithm in this topic is more accurate than Newman’s.

Key words: community structure, signal process, hierarchical clustering, modularity

摘要: 社团是社会网络的一个重要特征,社团发现是近年来研究的热点问题之一。通过在复杂网络上传递信号,获得各节点对网络的影响向量,从而把网络中节点的拓扑性质转化为代数空间上向量的几何关系,然后用结合模块度的层次聚类挖掘社会网络中的社团结构。该算法优点是不需要预先知道社团的数量或社团内节点的数量,用Zachary空手道俱乐部网络、大学足球赛网络以及海豚关系网络的数据进行验证,该算法划分的社团准确性超过了Newman的结论。

关键词: 社团结构, 信号传递, 层次聚类, 模块度

CLC Number: