电子学报 ›› 2016, Vol. 44 ›› Issue (3): 587-594.DOI: 10.3969/j.issn.0372-2112.2016.03.014

• 学术论文 • 上一篇    下一篇

基于隶属度的社会化网络重叠社区发现及动态集群演化分析

国琳1,2, 左万利1,2, 彭涛1,2   

  1. 1. 吉林大学计算科学与技术学院, 吉林长春 130012;
    2. 吉林大学符号计算与知识工程教育部重点实验室, 吉林长春 130012
  • 收稿日期:2014-06-07 修回日期:2014-10-15 出版日期:2016-03-25
    • 通讯作者:
    • 左万利
    • 作者简介:
    • 国琳 女,1987年8月生于吉林吉林,2013年至今于吉林大学计算机学院攻读博士学位,从事社会化网络、知识挖据及搜索引擎有关研究. E-mail:guolin13@mails.jlu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.60973040); 国家自然科学青年基金 (No.61300148); 吉林省重点科技攻关项目基金 (No.20130206051GX); 吉林省科技计划青年科研基金 (No.20130522112JH); 中国博士后基金项目 (No.2012M510879); 吉林大学基本科研业务费科学前沿与交叉项目 (No.201103129)

Overlapping Community Detection and Dynamic Group Evolution Analysis Based on the Degree of Membership in Social Network

GUO Lin1,2, ZUO Wan-li1,2, PENG Tao1,2   

  1. 1. College of Computer Science and Technology of Jilin University, Changchun, Jilin 130012, China;
    2. Symbol Computation and Knowledge Engineer of Ministry of Education of Jilin University, Changchun, Jilin 130012, China
  • Received:2014-06-07 Revised:2014-10-15 Online:2016-03-25 Published:2016-03-25
    • Supported by:
    • National Natural Science Foundation of China (No.60973040); Youth Fund of National Natural Science Foundation of China (No.61300148); Key Technology Research and Development Program of Jilin Province (No.20130206051GX); Youth Research Fund of Science and Technology Program of Jilin Province (No.20130522112JH); Post-doctoral Foundation of China (No.2012M510879); Fundamental Fund for Science Frontier and Cross Project of Jilin University (No.201103129)

摘要:

社会化网络中节点的复合属性可能为临时或过时状态,并且节点拥有一定能力维持固有状态,所以不可单纯依据新增数据或节点现有特征确定社区划分.本文提出可重叠社区发现算法及集群动态更新方案,根据网络历史数据分析节点对原始集群的隶属程度,并结合新增数据确定节点变化趋势,实现网络结构分析及社区动态更新.本文分别在不同数据集中测试聚类效果,实验结果证明算法既保持对新增数据的敏感度,也防止了节点短暂特征或节点维持固有状态的能力对划分结果的负面影响.

关键词: 社区发现, 社会化网络, 聚类, 重叠社区, 自适应算法

Abstract:

The complex social attributes of nodes in the network have a certain ability to maintain the former state,so it is inappropriate to determine community division merely based on newly added data.This paper proposes an overlapping community detection algorithm and dynamic cluster update strategy,which,by fully analyzing historical network data to compute the degree of nodes belonging to communities,determines the evolution tendency of nodes through incorporating incremental data to analyze the structure of the network and update the division results automatically.Experiments on several typical datasets demonstrate that the algorithm not only ensures the sensitivity to incremental data,but also avoids the negative effect of temporary features in maintaining intrinsic states on the clustering results.

Key words: community detection, social internet, clustering, overlapping community, adaptive algorithm

中图分类号: