基于密度峰值的重叠社区发现算法  

Overlapping community detection method based on density peaks

作　　者：时小虎[1,2,3,4] 冯国香[1,2] 李牧[1,2] 李瑛[1,2] 吴春国[1,2,3,4] SHI Xiao-hu FENG Guo-xiang LI Mu LI Ying WU Chun-guo(College of Computer Science and Technology, Jilin University, Changchun 130012, China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, China Department of Computer Science and Technology, Zhuhai College of Jilin University, Zhuhai 519041, China Zhuhai Laboratory of Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Zhuhai College of Jilin University, Zhuhai 519041, China)

机构地区：[1]吉林大学计算机科学与技术学院,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,长春130012 [3]吉林大学珠海学院计算机科学与技术系,广东珠海519041 [4]吉林大学珠海学院符号计算与知识工程教育部重点实验室珠海分实验室,广东珠海519041

出　　处：《吉林大学学报（工学版）》2017年第1期242-248,共7页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家自然科学基金项目(61373050);吉林省科技发展计划项目(20130101070JC;20130522118JH;20130206003SF);在线教育研究基金(全通教育)项目(2017YB129);广东省优势重点学科建设项目;珠海市优势学科建设项目

摘　　要：根据基于快速搜索和发现密度峰值的聚类方法的思想,提出了基于密度峰值的重叠社区发现算法。首先定义新的距离矩阵算法,克服了邻接矩阵元素为整数的缺陷。然后用概率形式刻画每个节点属于不同类别的可能性,从而实现了重叠社区的划分。基于真实网络的实验结果验证了本文算法的可行性和有效性。Based on the idea of clustering method by fast search and find of density peaks,we put forward an overlapping community detection method based on density peaks.In the proposed algorithm,first a new distance matrix is defined to overcome the defects of adjacency matrix being filled with integers.Then,the probability form is used to depict the possibility of each node that belongs to different clusters.The experimental results on real network show that the proposed method performs very well.

关 键 词：计算机应用 复杂网络 社区发现 重叠社区 小世界特性 模块度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]