基于HMM的动态社会网络社团发现算法

发布时间：2018-12-11 00:33

【摘要】：随着互联网的不断发展,大多数社会网络已逐渐显示出动态特性,动态社会网络社团分析对理解现实生活中社会网络结构和功能具有非常重要的意义.针对动态社会网络中的社团发现问题,提出一种基于隐Markov模型(hidden Markov model,HMM)的HMM_DC算法.该算法考虑到社会网络的动态特性,结合历史信息,将社团发现转化为求解隐马尔可夫模型中的最优状态序列问题,将网络中的社团结构和节点信息分别采用状态链和观察链表示,在无须指定额外参数的情况下实现动态网络的社团结构发现.最后,利用该算法和其他算法对VAST数据集、ENRON数据集和Facebook social network数据集进行实验仿真.仿真结果表明:该算法能够快速、准确地发现真实动态网络中的社团,其模块度Q值和互信息NMI值有很大提升.
[Abstract]:With the continuous development of the Internet, most social networks have gradually shown dynamic characteristics. Community analysis of dynamic social networks is of great significance to understand the social network structure and functions in real life. To solve the community discovery problem in dynamic social networks, a HMM_DC algorithm based on hidden Markov model (hidden Markov model,HMM) is proposed. Considering the dynamic characteristics of social network and historical information, the algorithm transforms community discovery into solving the optimal state sequence problem in hidden Markov model. The community structure and node information in the network are represented by the state chain and the observation chain respectively. The community structure discovery of the dynamic network is realized without specifying additional parameters. Finally, the VAST dataset, ENRON dataset and Facebook social network dataset are simulated by using the algorithm and other algorithms. The simulation results show that the algorithm can quickly and accurately find the communities in the real dynamic network, and its module degree Q value and mutual information NMI value are greatly improved.
【作者单位】：国家数字交换系统工程技术研究中心;中华人民共和国科学技术部信息中心;
【基金】：国家“八六三”高技术研究发展计划基金项目(2015AA016102)~~
【分类号】：TP301.6

【相似文献】