基于多属性融合策略的复杂网络社团划分算法  

Community Detection Algorithm Based on Multi-attribute Fusion Strategy

作　　者：乔虹 田玉玲 马建芬 QIAO Hong;TIAN Yu-ling;MA Jian-fen(College of Computer Science and Technology,Taiyuan University of Technology,Jinzhong 030600,China)

机构地区：[1]太原理工大学信息与计算机学院,晋中030600

出　　处：《科学技术与工程》2018年第32期51-57,共7页Science Technology and Engineering

基　　金：国家自然科学基金(61472271)资助

摘　　要：为解决目前社团划分算法依赖于单一度量方法,划分结果不够准确,无法适应不同类型的网络划分需求的问题,通过一种多属性融合策略方法研究社团划分。该算法首先基于节点连接,综合度量了多个影响社团划分的属性,并引入模块度概念确定各属性融合的权重,为划分增加了客观的参考信息,从而提高划分准确率;其次,针对传统划分方法迭代次数过多、效率低的问题,利用人工免疫网络内在的全局并行搜索能力实现对社团核心节点的快速寻优,并提出动态算子、免疫检测因子和反向学习机制对人工免疫网络的收敛速度和局部最优问题加以改进,从而提高寻优效率,缩短算法执行时间。最后,在三个经典真实数据集(Zachary、Dolphin、College Football)上进行实验,并将结果与经典算法对比。结果表明,该算法能适应不同的网络,且在较短的执行时间里实现更加精确的划分。可见,相比传统算法,本文算法具有更高的划分效率。Most community detection methods rely on a single method to measure the connection strength between nodes,and lead to not perfect effects,and they are unable to adapt to different types of Network,either.In order to solve above problems,a novel community detection algorithm based on multi-attribute fusion strategy was used to investigate community detection.Firstly,from the view of multi-attribute fusion strategy,the proposed algorithm comprehensively measured multiple factors which affecting community detection,and introduced the modularity to determine weights of attributes to get more objective reference for detection and achieve more accurate effects.Secondly,to decrease excessive iteration times and improve efficiency of traditional methods,the algorithm further used artificial immune network with powerful parallel computing ability to find the core nodes of the community rapidly,and improved convergence rate and solved the local optimal problem by using dynamic operators,immune detection operators and opposition-learning.The experimental results show that the proposed algorithm gets better results on three different datasets of Zachary,Dolphin and College Football compared with traditional methods.It is concluded that the proposed algorithm achieves more accurate division effects in shorter time and has greater efficiency.

关 键 词：社团结构 社团划分 多属性 人工免疫网络 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]