摘要
社区发现是复杂网络分析的重要任务之一。现有的社区发现方法大多面向单层网络,对现实世界中广泛存在的多层网络中的社区发现研究则稍显不足。首先给出了各种多层网络的定义,并对比了各网络在节点对齐、层间边和层间耦合方面的特点,接着介绍了各类传统的单层网络社区发现方法;在此基础上,深入调研了多层网络社区发现方法,将其大致分为基于聚合的方法和基于扩展的方法,对各方法在实现机制、优势、局限性、适用网络、复杂性等方面进行了分析比较;通过在现实和模拟数据集上的实验,进一步比较了部分代表性方法在模块度、标准化互信息和调整兰德指数等指标上的性能;最后对多层网络社区发现的工作进行了总结和展望。
Community detection is one of the most important tasks of complex network analysis.Most of the existing community detection methods are oriented to single-layer networks,and the research on community detection in multi-layer networks widely existing in the real world is slightly insufficient.This paper first presents the definition of various multi-layer networks,compares the characteristics of each network in terms of node alignment,inter-layer edges and inter-layer coupling,and then introduces various traditional single-layer network community discovery methods.On this basis,this paper deeply surveys the multi-layer network community detection methods,which are roughly divided into aggregation-based methods and extension-based methods.The methods are analyzed and compared in terms of mechanisms,advantages,limitations,application network,complexity,etc.Experiments are conducted on both real and simulated datasets,comparing the performance of several representative methods on indicators such as modularity,normalized mutual information,and adjustment of the Rand index.The work of multilayer network community detection is summarized and prospected.
作者
陈可佳
陈利明
吴桐
CHEN Kejia;CHEN Liming;WU Tong(School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China;Jiangsu Key Laboratory of Big Data Security&Intelligent Processing,Nanjing 210023,China)
出处
《计算机科学与探索》
CSCD
北大核心
2020年第11期1801-1812,共12页
Journal of Frontiers of Computer Science and Technology
基金
国家自然科学基金(Nos.61772284,61603197)。
关键词
社区发现
多层网络
模块度
复杂网络分析
community detection
multi-layer network
modularity
complex network analysis
