基于邻域信息的社区发现方法  

Community detection based on neighborhood information

作　　者：韩路[1] 张海[1] 

机构地区：[1]西北大学数学学院,陕西西安710127

出　　处：《纯粹数学与应用数学》2015年第1期85-92,共8页Pure and Applied Mathematics

基　　金：国家自然科学基金(11171272)

摘　　要：考虑含有节点邻域信息的新模块度函数的社区发现方法和最优分组下标度参数的选择问题,通过谱松弛方法求解模块度函数的最大化问题,最终利用新算法快速求解,并通过真实网络数据验证算法能更好的发现社区.Community detection based on modularity is a widely used method, but it does not use neighborhood information of nodes, then it fails to be a good representation of real-world community structure. A new modularity with neighborhood information could detect the community structure of real-world networks, but it didn＇t show parameter selection of the best community division. The paper aimed at the maximization and parameter selection of the new modularity with neighborhood information, then reformulate the maximization as a spectral relaxation issue. Finally, we solve the problem by a new bisection spectral algorithm and prove the effectiveness of our algorithm by experimental results.

关 键 词：模块度函数 邻域信息 谱方法 

分 类 号：O233[理学—运筹学与控制论] TP391.41[理学—数学]