动态属性网络的语义社区发现及演化分析方法  

A Community Discovery and Evolution Analysis Method for Dynamic Attributed Networks

作　　者：贺超波[1,2] 成其伟 程俊伟 杨佳琦 程颢 汤庸 HE Chao-bo;CHENG Qi-wei;CHENG Jun-wei;YANG Jia-qi;CHENG Hao;TANG Yong(School of Computer Science,South China Normal University,Guangzhou,Guangdong 510631,China;Pazhou Lab,Guangzhou,Guangdong 510335,China;Vivo Mobile Communication Co.,Ltd.,Dongguan,Guangdong 523859,China)

机构地区：[1]华南师范大学计算机学院,广东广州510631 [2]琶洲实验室,广东广州510335 [3]维沃移动通信有限公司,广东东莞523859

出　　处：《电子学报》2024年第11期3757-3768,共12页Acta Electronica Sinica

基　　金：国家自然科学基金(No.62077045)。

摘　　要：动态属性网络的语义社区发现及演化分析具有重要研究价值,其包含动态社区发现、社区语义解释及社区演化分析三个任务,但现有方法均难以同时实现.针对该问题,提出一种基于联合非负矩阵分解的方法DANNMF(NMF for Dynamic Attributed Networks).DAN-NMF可以统一集成网络拓扑结构信息、节点属性信息及社区演化平滑约束信息,并利用最大最小化优化框架推导相关因子矩阵的迭代更新规则,从而可以直接获得动态社区发现、社区语义解释及社区演化分析结果.在人工合成和真实的动态属性网络进行大量相关实验,结果表明DAN-NMF比最优的基准方法在准确性指标上至少提高了7.3%.此外,在真实动态属性网络上的相关数据分析结果也表明DAN-NMF能够有效地发现动态社区的演化模式,并提供丰富的社区语义解释.The topic of semantic community discovery and evolution analysis in dynamic attributed networks has im⁃portant research value.It needs to simultaneously accomplish the tasks of dynamic community discovery,community se⁃mantic interpretation and community evolution analysis,but existing methods are difficult to achieve this goal.In view of this,this paper proposes a method DAN-NMF(NMF for Dynamic Attributed Networks)based on joint nonnegative matrix factorization.DAN-NMF can uniformly integrate network topology information,attribute information and smooth con⁃straint information from community evolution,and derive iterative update rules of the related factor matrices using the ma⁃jorization-minimization optimization framework,which helps it to directly obtain the results of dynamic community discov⁃ery,community semantic interpretation and community evolution analysis.Extensive experiments are conducted on multi⁃ple synthetic and real-world dynamic attributed networks.The results show that DAN-NMF has improved by at least 7.3%in term of accuracy metric,compared to the optimal baseline.Moreover,the data analysis results on real-world dynamic at⁃tributed networks also demonstrate that DAN-NMF can effectively discover the evolution patterns of dynamic communities and provide rich community semantic interpretations.

关 键 词：动态属性网络 动态社区发现 社区语义解释 社区演化分析 非负矩阵分解 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]