基于结构增强和深度聚类的网络群体识别

doi:10.3969/j.issn.0255-8297.2026.01.001

摘要/Abstract

摘要： 社交网络群体识别在信息传播、推荐与广告等领域具有重要的研究意义与应用价值。但现有方法在特征融合、稀疏图建模及多源信息利用上仍存在不足。为此，提出一种基于结构增强与深度聚类（structure-enhanced deep clustering,SDC）的网络群体识别模型，包含4个关键模块：首先，网络拓扑增强模块通过建模节点二阶相似性生成增强邻接矩阵，缓解稀疏社交网络的高阶关系缺失；其次，多视图特征融合模块在节点级动态融合节点属性特征与拓扑特征，在图级整合原始图与增强图的语义信息；再次，多源分布融合聚类模块在分布级利用可学习权重集成不同特征空间的聚类信息，平衡局部拓扑与全局语义；最后，双重自监督模块通过KL散度（Kullback-Leibler divergence）对齐、节点重构与相似性约束进行优化。实验表明，相较于主流基线方法，SDC网络群体识别模型在3个基准数据集上的ACC、NMI、ARI、F1指标平均提升了3.80%、9.09%、11.21%和7.43%。在Facebook动态交互数据上的仿真也验证了SDC网络群体识别模型捕捉社区结构演化的能力。

关键词: 社交网络, 群体识别, 结构增强, 深度聚类, 注意力机制

Abstract: Community detection in social networks is important for applications such as information diffusion, recommendation, and advertising. However, existing methods still face challenges in feature fusion, sparse graph modeling, and multi-source information utilization. To address these issues, this paper proposed a structure-enhanced deep clustering (SDC) model for network community detection, which consisted of four key modules. First, the topology enhancement module built an enhanced adjacency matrix by modeling second-order similarity between nodes, which alleviated the problem of missing high-order relations in sparse social networks. Second, the multi-view feature fusion module dynamically fused node attributes and topology features at the node level, and integrated semantic information from both the original and enhanced graphs at the graph level. Third, the multi-source distribution fusion clustering module used learnable weights to integrate clustering information from different feature spaces at the distribution level, balancing local topology and global semantics. Finally, the dual self-supervised module optimized the model through Kullback-Leibler (KL) divergence alignment, node reconstruction, and similarity constraints. Experiments show that compared with mainstream baseline methods, SDC model improves ACC, NMI, ARI, and F1 by an average of 3.80%, 9.09%, 11.21%, and 7.43%, respectively on the three benchmark datasets. Simulations based on Facebook interaction data also demonstrate the ability of SDC model to capture community structure evolution.

Key words: social network, community detection, structure-enhanced, deep clustering, attention mechanism

中图分类号:

TP391.41

李永桢, 马涪元, 马世旋, 王钰涵, 王英. 基于结构增强和深度聚类的网络群体识别[J]. 应用科学学报, 2026, 44(1): 1-20.

LI Yongzhen, MA Fuyuan, MA Shixuan, WANG Yuhan, WANG Ying. Network Community Detection Based on Structure-Enhanced Deep Clustering[J]. Journal of Applied Sciences, 2026, 44(1): 1-20.

参考文献

[1] Barnes J A. Class and committees in a Norwegian Island parish [J]. Human Relations, 1954, 7(1): 39-58.
[2] Centola D. The spread of behavior in an online social network experiment [J]. Science, 2010, 329(5996): 1194-1197.
[3] Leskovec J, Adamic L A, Huberman B A. The dynamics of viral marketing [J]. ACM Transactions on the Web, 2007, 1(1): 5.
[4] Song X D, Tseng B L, Lin C Y, et al. Personalized recommendation driven by information flow [C]//29th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2006: 509-516.
[5] Tajeuna E G, Bouguessa M, Wang S R. Modeling and predicting community structure changes in time-evolving social networks [J]. IEEE Transactions on Knowledge and Data Engineering, 2019, 31(6): 1166-1180.
[6] Chen X Q, Zhou F, Zhang K P, et al. Information diffusion prediction via recurrent cascades convolution [C]//IEEE 35th International Conference on Data Engineering (ICDE), 2019: 770- 781.
[7] He X N, Gao M, Kan M Y, et al. BiRank: towards ranking on bipartite graphs [J]. IEEE Transactions on Knowledge and Data Engineering, 2017, 29(1): 57-71.
[8] Wang X, Nie L Q, Song X M, et al. Unifying virtual and physical worlds: learning toward local and global consistency [J]. ACM Transactions on Information Systems, 2018, 36(1): 1-26.
[9] Blondel V D, Guillaume J L, Lambiotte R, et al. Fast unfolding of communities in large networks [J]. Journal of Statistical Mechanics: Theory and Experiment, 2008, 10: P10008.
[10] Šubelj L. Label propagation for clustering [J]. Advances in Network Clustering and Blockmodeling, 2019: 121-150.
[11] Girvan M, Newman M E J. Community structure in social and biological networks [J]. Proceedings of the National Academy of Sciences of the United States of America, 2002, 99(12): 7821-7826.
[12] Xiao Q, Du S, Zhang K, et al. Adaptive sparse graph learning for multi-view spectral clustering [J]. Applied Intelligence, 2023, 53(12): 14855-14875.
[13] Pan S, Hu R, Long G, et al. Adversarially regularized graph autoencoder for graph embedding [DB/OL]. (2018-02-13) [2025-08-05]. https://arxiv.org/abs/1802.04407.
[14] Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks [DB/OL]. (2016-09-09) [2025-08-05]. https://arxiv.org/abs/1609.02907.
[15] Xie J, Girdhick R, Farhadi A. Unsupervised deep embedding for clustering analysis [C]//International Conference on Machine Learning, 2016: 478-487.
[16] Airoldi E M, Blei D, Fienberg S, et al. Mixed membership stochastic blockmodels [J]. Journal of Machine Learning Research, 2008(9): 1981-2014.
[17] Kipf T N, Welling M. Variational graph auto-encoders [DB/OL]. (2016-11-21) [2025-08-05]. https://arxiv.org/abs/1611.07308.
[18] Bo D Y, Wang X, Shi C, et al. Structural deep clustering network [C]//Web Conference, 2020: 1400-1410.
[19] Wang C, Pan S, Hu R, et al. Attributed graph clustering: a deep attentional embedding approach [DB/OL]. (2019-06-15) [2025-08-05]. https://arxiv.org/abs/1906.06532.
[20] Wu Y L, Fu Y, Xu J W, et al. Heterogeneous question answering community detection based on graph neural network [J]. Information Sciences, 2023, 621: 652-671.
[21] Costa A R, Ralha C G. AC2CD: an actor–critic architecture for community detection in dynamic social networks [J]. Knowledge-Based Systems, 2023, 261: 110202.
[22] Stephan L, Zhu Y Z. Sparse random hypergraphs: non-backtracking spectra and community detection [J]. Information and Inference: A Journal of the IMA, 2024, 13(1): iaae004.
[23] Yang X H, Liu Y, Zhou S H, et al. Cluster-guided contrastive graph clustering network [J]. AAAI Conference on Artificial Intelligence, 2023, 37(9): 10834-10842.
[24] Reihanian A, Feizi-Derakhshi M R, Aghdasi H S. An enhanced multi-objective biogeography-based optimization for overlapping community detection in social networks with node attributes [J]. Information Sciences, 2023, 622: 903-929.
[25] Liu Y, Yang X H, Zhou S H, et al. Simple contrastive graph clustering [J]. IEEE Transactions on Neural Networks and Learning Systems, 2024, 35(10): 13789-13800.
[26] Hajiveiseh A, Seyedi S A, Tab F A. Deep asymmetric nonnegative matrix factorization for graph clustering [J]. Pattern Recognition, 2024, 148: 110179.
[27] Li H J, Feng Y H, Xia C Y, et al. Overlapping graph clustering in attributed networks via generalized cluster potential game [J]. ACM Transactions on Knowledge Discovery from Data, 2024, 18(1): 1-26.
[28] Berahmand K, Mohammadi M, Sheikhpour R, et al. WSNMF: weighted symmetric nonnegative matrix factorization for attributed graph clustering [J]. Neurocomputing, 2024, 566: 127041.
[29] Peng X, Cheng J R, Tang X Y, et al. Dual contrastive learning network for graph clustering [J]. IEEE Transactions on Neural Networks and Learning Systems, 2024, 35(8): 10846-10856.
[30] Mcpherson M, Smith-Lovin L, Cook J M. Birds of a feather: homophily in social networks [J]. Annual Review of Sociology, 2001, 27: 415-444.
[31] Pirasteh P, Hwang D, Jung J J. Exploiting matrix factorization to asymmetric user similarities in recommendation systems [J]. Knowledge-Based Systems, 2015, 83: 51-57.
[32] Student. The probable error of a mean [J]. Biometrika, 1908, 6(1): 1-25.
[33] Kullbak S, Leibler R A. On information and sufficiency [J]. The Annals of Mathematical Statistics, 1951, 22(1): 79-86.
[34] Peng Z H, Liu H, Jia Y H, et al. Attention-driven graph clustering network [C]//29th ACM International Conference on Multimedia, 2021: 935-943.
[35] Tu W X, Zhou S H, Liu X W, et al. Deep fusion clustering network [J]. AAAI Conference on Artificial Intelligence, 2021, 35(11): 9978-9987.
[36] Li X, Li D, Jin R M, et al. Deep graph clustering with random-walk based scalable learning [C]//IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), 2022: 88-95.
[37] Fanseu K B, Zhang L, Ma K L, et al. GRACE: a general graph convolution framework for attributed graph clustering [J]. ACM Transactions on Knowledge Discovery from Data, 2023, 17(3): 1-31.