复杂网络社团检测可视化仿真平台设计与开发  

Design and Development of Visual and Emulational Platform of Complex Network Community Detection

作　　者：卢钊杰 陈珂[1] 蓝鼎栋 吴镇佳 张良均 

机构地区：[1]广东石油化工学院计算机科学与技术系,广东茂名525000 [2]广州太普信息技术有限公司,广东广州510663

出　　处：《广东石油化工学院学报》2015年第6期33-37,共5页Journal of Guangdong University of Petrochemical Technology

基　　金：国家级大学生创新创业训练计划项目(201411656017);广东省科技计划项目(2012B0101100037);广东省高等学校科技创新项目(2013kjcx0132)

摘　　要：有效发现与理解社团结构是复杂网络研究中非常重要的问题。此文分析与比较了传统的GN算法、Fast-Newman算法和CNM算法,基于JUNG框架设计与开发了社团检测可视化平台。该平台首先实现了上述三种算法,并与此文提出的局部社团检测算法(CDAMDL算法)在标准网络数据集上进行了分析测试。实验表明,在检测效果,计算耗时等方面,CDAMDL算法均优于传统算法。同时该平台提供了透明、直观的社团检测结果和评价,为复杂网络社团结构分析提供了依据。How to effectively find community structure in complex network and understand the whole network structure and function is a very important problem. This article analyses and compares the traditional algorithm, GN algorithm, Fast - Newman algorithm and CNM algo- rithm, designing the community detection visual platform based on JUNG framework. The platform firstly achieves three kinds of algorithms mentioned above, and then they are compared with the local community detection algorithm （CDAMDL algorithm） in a lot of experiments on standard web data sets. The experimental results show that, CDAMDL algorithm is better than the traditional algorithm in calculation time consuming. At the same time, the platform provides a transparent, intuitive association testing results and evaluation, and provides the basis for complex network community structure analysis.

关 键 词：社团检测算法 局部检测 可视化 JUNG 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]