Multi-level neighborhood contrastive attribute graph clustering based on adaptive learning

doi:10.11772/j.issn.1001-9081.2025050647

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (6): 1836-1843.DOI: 10.11772/j.issn.1001-9081.2025050647

• Data science and technology • Previous Articles

Multi-level neighborhood contrastive attribute graph clustering based on adaptive learning

Jinghong WANG¹^,²^,³^,⁴^,⁵, Xiao CHEN¹, Yingmei MA²^,⁵(), Bi LI⁶, Jusheng MI⁷, Wei WANG¹

^1.College of Computer and Cyber Security，Hebei Normal University，Shijiazhuang Hebei 050024，China
^2.College of Artificial Intelligence，Hebei University of Engineering Science，Shijiazhuang Hebei 050091，China
^3.Hebei Key Laboratory of Network and Information Security （Hebei Normal University），Shijiazhuang Hebei 050024，China
^4.Hebei Provincial Engineering Research Center for Supply Chain Big Data Analytics and Data Security （Hebei Normal University），Shijiazhuang Hebei 050024，China
^5.Shijiazhuang Intelligent IoT Technology Innovation Center，Shijiazhuang Hebei 050299，China
^6.School of Business，Hebei Normal University，Shijiazhuang Hebei 050024，China
^7.School of Mathematical Sciences，Hebei Normal University，Shijiazhuang Hebei 050024，China

Received:2025-06-12 Revised:2025-08-16 Accepted:2025-09-01 Online:2025-09-15 Published:2026-06-10
Contact: Yingmei MA
About author:WANG Jinghong， born in 1967， Ph. D.， professor. Her research interests include artificial intelligence， data mining.
CHEN Xiao， born in 2000， M. S. candidate. Her research interests include data mining， graph representation learning.
LI Bi， born in 1966， M. S.， associate professor. His research interests include fuzzy systems.
MI Jusheng， born in 1966， Ph. D.， professor. His research interests include mathematical foundation of artificial intelligence， rough set theory.
WANG Wei， born in 1982， Ph. D.， associate professor. His research interests include computer vision.
First author contact:MA Yingmei， born in 1989， M. S.， lecturer. Her research interests include big data analysis.
Supported by:
Natural Science Foundation of Hebei Province(F2024205028);Post-graduate’s Innovation Fund Project of Hebei Province(CXZZSS2025049);Key Development Fund of Hebei Normal University(L2024ZD06)

基于自适应学习的多层次邻域对比属性图聚类

王静红¹^,²^,³^,⁴^,⁵, 陈潇¹, 马迎梅²^,⁵(), 李笔⁶, 米据生⁷, 王威¹

^1.河北师范大学计算机与网络空间安全学院，石家庄 050024
^2.河北工程技术学院人工智能学院，石家庄 050091
^3.河北省网络与信息安全重点实验室（河北师范大学），石家庄 050024
^4.供应链大数据分析与数据安全河北省工程研究中心（河北师范大学），石家庄 050024
^5.石家庄市智能物联技术创新中心，石家庄 050299
^6.河北师范大学商学院，石家庄 050024
^7.河北师范大学数学科学学院，石家庄 050024

通讯作者: 马迎梅
作者简介:王静红（1967—），女，河北石家庄人，教授，博士，CCF会员，主要研究方向：人工智能、数据挖掘
陈潇（2000—），女，河北沧州人，硕士研究生，CCF会员，主要研究方向：数据挖掘、图表示学习
李笔（1966—），男，湖南湘乡人，副教授，硕士，主要研究方向：模糊系统
米据生（1966—），男，河北宁晋人，教授，博士，主要研究方向：人工智能的数学基础、粗糙集理论
王威（1982—），男，河北邯郸人，副教授，博士，主要研究方向：计算机视觉。
第一联系人：马迎梅（1989—），女，河北石家庄人，讲师，硕士，主要研究方向：大数据分析
基金资助:
河北省自然科学基金资助项目(F2024205028);河北省研究生创新资助项目(CXZZSS2025049);河北师范大学重点发展基金资助项目(L2024ZD06)

Abstract

Abstract:

Recently， deep graph clustering methods have outstanding performance in graph clustering studies. However， most existing deep graph clustering methods are based on the auto-encoder framework， and are vulnerable to reconstruction strategies and graph enhancement strategies. Therefore， a deep graph clustering method based on contrastive learning was proposed， namely Multi-level Neighborhood Contrastive attribute Graph Clustering based on adaptive learning （MNCGC）. Firstly， a dual masking strategy was designed to generate an adaptive augmented graph， which combined the node importance to generate edge weights， that is， edge masking probabilities， and a fixed masking probability was set for node features for node feature masking， so as to remove redundant information in the graph and provide rich sample pairs for neighborhood contrastive learning. Then， the edge weights were introduced into the neighborhood contrastive learning， so that the enhanced neighborhood contrastive learning was used to the original graph and the augmented graph at coding level and projection level， thereby emphasizing the local information learning and the global high-level semantic information learning. Finally， self-supervised clustering and code level representation were used to promote each other， thereby further improving the clustering effect. Experimental results on three benchmark datasets including Cora， CiteSeer and PubMed show that compared with fourteen advanced methods， MNCGC method achieves optimal values in most cases across four indicators： accuracy， Normalized Mutual Information （NMI）， Adjusted Rand Index （ARI） and F1-score， fully verifying the effectiveness of the proposed method.

Key words: attribute graph clustering, adaptive learning, contrastive learning, self-supervised clustering, deep graph clustering

摘要：

在近期的图聚类研究中，深度图聚类方法表现优异；然而，现有的深度图聚类方法多数基于自编码器框架，容易受到重建策略和图增强策略的影响。因此，提出一种基于对比学习的深度图聚类方法——基于自适应学习的多层次邻域对比属性图聚类（MNCGC）。首先，设计一种双重掩蔽策略生成自适应增广图，结合节点重要性生成边权重，即边屏蔽概率，并对节点特征设置固定屏蔽概率，进行节点特征屏蔽，去除图中的冗余信息，为邻域对比学习提供丰富的样本对；其次，将边权重引入邻域对比学习，以对原图和增广图使用加强邻域对比学习并把它们应用于编码级和投影级，从而强调局部信息学习和全局高级语义信息学习；最后，采用自监督聚类与编码级表示相互促进，进一步提升聚类效果。在3个基准数据集（Cora、CiteSeer和全文PubMed）上的实验结果表明，相较于14种先进方法，MNCGC方法的准确率、标准互信息（NMI）、调整兰德指数（ARI）以及F1分数在大多数情况下都取得了最优值，充分验证了所提方法的有效性。

关键词: 属性图聚类, 自适应学习, 对比学习, 自监督聚类, 深度图聚类

CLC Number:

TP391.4

Jinghong WANG, Xiao CHEN, Yingmei MA, Bi LI, Jusheng MI, Wei WANG. Multi-level neighborhood contrastive attribute graph clustering based on adaptive learning[J]. Journal of Computer Applications, 2026, 46(6): 1836-1843.

王静红, 陈潇, 马迎梅, 李笔, 米据生, 王威. 基于自适应学习的多层次邻域对比属性图聚类[J]. 《计算机应用》唯一官方网站, 2026, 46(6): 1836-1843.

Figures/Tables 11

Tab. 1 Symbol definition

符号	含义	符号	含义
$G$	属性图	$H$	编码表示矩阵
$V$	节点集	$M$	投影表示矩阵
$E$	边集	$T$	目标分布更新间隔
$X$	节点特征矩阵	$τ$	温度参数
$D$	度矩阵	$e c o$	编码级损失超参数
n	节点数	pro	投影级损失超参数
$K$	集群数	$Q$	生成分布函数
$A$	邻接矩阵	$P$	目标分布函数
$A'$	增广邻接矩阵	$μ$	聚类中心
$Z$	嵌入矩阵	$y'$	预测标签
$B$	边嵌入向量	$α$	对比学习优化超参数
$W$	边权重矩阵	$β$	自监督聚类优化超参数

Tab. 1 Symbol definition

符号	含义	符号	含义
$G$	属性图	$H$	编码表示矩阵
$V$	节点集	$M$	投影表示矩阵
$E$	边集	$T$	目标分布更新间隔
$X$	节点特征矩阵	$τ$	温度参数
$D$	度矩阵	$e c o$	编码级损失超参数
n	节点数	pro	投影级损失超参数
$K$	集群数	$Q$	生成分布函数
$A$	邻接矩阵	$P$	目标分布函数
$A'$	增广邻接矩阵	$μ$	聚类中心
$Z$	嵌入矩阵	$y'$	预测标签
$B$	边嵌入向量	$α$	对比学习优化超参数
$W$	边权重矩阵	$β$	自监督聚类优化超参数

Fig. 1 Framework of MNCGC method

Tab. 2 Dataset information

数据集	节点数	边数	特征维度	标签数
Cora	2 708	5 429	1 433	7
CiteSeer	3 327	4 732	3 703	6
PubMed	19 717	44 338	500	3

Tab. 3 Hyperparameter setting

超参数	含义	值
$e p o c h$	预训练次数	120
$l r$	学习率	0.000 1
$e p o c h'$	正式训练次数	50
$e c o$	编码级损失超参数	1
$p r o$	投影级损失超参数	0.1
$α$	对比学习优化超参数	1
$β$	自监督聚类优化超参数	1

Tab. 3 Hyperparameter setting

超参数	含义	值
$e p o c h$	预训练次数	120
$l r$	学习率	0.000 1
$e p o c h'$	正式训练次数	50
$e c o$	编码级损失超参数	1
$p r o$	投影级损失超参数	0.1
$α$	对比学习优化超参数	1
$β$	自监督聚类优化超参数	1

Tab. 4 Comparison of clustering performance of methods on three experimental datasets

方法	Cora				CiteSeer				PubMed
方法	ACC	F1	NMI	ARI	ACC	F1	NMI	ARI	ACC	F1	NMI	ARI
GRACE	60.22	58.01	47.11	35.91	62.13	62.04	37.90	37.15	Out of memory
SDCN	6.67	27.41	18.73	8.29	64.29	59.32	36.49	38.17	62.32	61.53	23.04	21.64
AGCN	38.74	28.59	18.55	9.62	67.54	62.46	39.12	41.85	62.44	60.52	23.94	21.67
GCA	56.33	51.13	46.36	30.16	64.47	60.03	39.57	39.45	66.67	66.18	33.11	28.33
GC-VGE	70.68	69.48	53.57	48.15	66.61	63.39	40.91	41.52	68.18	66.88	29.71	29.76
DNENC-con	68.30	65.90	51.20	47.70	69.20	63.90	42.60	44.90	67.70	67.50	27.50	27.80
DNENC-att	70.40	68.20	52.80	49.60	67.20	63.60	39.70	41.00	67.10	65.90	26.60	27.80
AGGDC	74.60	—	59.50	52.20	69.60	—	45.30	45.80	61.90	—	33.60	29.50
ASP	65.81	60.58	55.99	46.78	68.66	64.20	43.78	44.83	Out of memory
CCGC	74.36	71.28	56.32	52.13	69.35	62.21	43.47	44.12	62.47	61.38	27.72	25.67
CGC	75.15	66.22	56.90	55.27	69.31	64.74	43.61	42.14	67.43	67.14	33.07	32.27
CoCGC	72.51	68.71	56.15	49.49	70.15	62.88	44.58	45.29	—	—	—	—
AMGC	66.65	61.02	47.99	43.40	60.92	57.33	32.93	33.73	64.56	64.52	24.58	24.19
MPCCL	72.03	69.73	53.86	52.29	70.56	66.09	45.10	46.90	—	—	—	—
MNCGC	75.52	73.42	59.57	58.06	70.24	66.75	45.43	46.12	71.08	70.84	35.57	34.48

Fig. 2 Experimental results comparison of variant study

Fig. 3 Impact of hyperparameters of multi-level comparison module

Fig. 4 Impact of hyperparameters of two modules

Tab. 5 Experimental results comparison of parameter sensitivity （ACC）

lr	Cora			CiteSeer			PubMed
lr	τ=0.5	τ=1.0	τ=1.5	τ=0.5	τ=1.0	τ=1.5	τ=0.5	τ=1.0	τ=1.5
0.000 1	72.51	72.71	72.35	67.48	68.82	68.67	67.65	69.11	68.04
0.000 3	74.68	75.52	73.79	68.51	70.24	67.32	68.47	70.14	68.71
0.000 5	73.06	73.39	73.01	66.53	69.33	66.58	68.29	71.08	69.24

Fig. 5 Impact of edge weight

Fig. 6 Convergence analysis on Cora dataset

References 30

[1]	LIU X， ZHANG F， HOU Z， et al. Self-supervised learning： generative or contrastive［J］. IEEE Transactions on Knowledge and Data Engineering， 2023， 35（1）： 857-876.
[2]	WU L， LIN H， TAN C， et al. Self-supervised learning on graphs： contrastive， generative， or predictive［J］. IEEE Transactions on Knowledge and Data Engineering， 2023， 35（4）： 4216-4235.
[3]	LIU J， CAO F， JING X， et al. Deep multi-view graph clustering network with weighting mechanism and collaborative training［J］. Expert Systems with Applications， 2024， 236： No.121298.
[4]	PAN S， HU R， LONG G， et al. Adversarially regularized graph autoencoder for graph embedding［C］// Proceedings of the 27th International Joint Conference on Artificial Intelligence. California： ijcai.org， 2018： 2609-2615.
[5]	PARK J， LEE M， CHANG H J， et al. Symmetric graph convolutional autoencoder for unsupervised graph representation learning［C］// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2019： 6518-6527.
[6]	WANG W， SUO X， WEI X， et al. HGATE： heterogeneous graph attention auto-encoders［J］. IEEE Transactions on Knowledge and Data Engineering， 2023， 35（4）： 3938-3951.
[7]	CHEN M， WANG B， LI X. Deep contrastive graph learning with clustering-oriented guidance［C］// Proceedings of the 38th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2024： 11364-11372.
[8]	LIU Y， YANG X， ZHOU S， et al. Simple contrastive graph clustering ［J］. IEEE Transactions on Neural Networks and Learning Systems， 2024， 35（10）： 13789-13800.
[9]	ZHONG H， WU J， CHEN C， et al. Graph contrastive clustering［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2021： 9204-9213.
[10]	安俊秀，柳源，杨林旺. 基于对比学习的深度聚类研究综述［J］. 微电子学与计算机， 2025， 42（7）： 1-10.
	AN J X， LIU Y， YANG L W. A review of deep clustering research based on comparative learning［J］. Microelectronics & Computer， 2025， 42（7）： 1-10.
[11]	SHEN X， SUN D， PAN S， et al. Neighbor contrastive learning on learnable graph augmentation［C］// Proceedings of the 37th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2023： 9782-9791.
[12]	ZHENG Y， JIA C， YU J. Attributed graph clustering under the contrastive mechanism with cluster-preserving augmentation［J］. Information Sciences， 2024， 681： No.121225.
[13]	YOU Y， CHEN T， SUI Y， et al. Graph contrastive learning with augmentations［C］// Proceedings of the 34th International Conference on Neural Information Processing Systems. Red Hook： Curran Associates Inc.， 2020： 5812-5823.
[14]	ZHU Y， XU Y， YU F， et al. Deep graph contrastive representation learning ［EB/OL］. ［2025-08-15］..
[15]	ZHU Y， XU Y， YU F， et al. Graph contrastive learning with adaptive augmentation ［C］// Proceedings of the Web Conference 2021. New York： ACM， 2021： 2069-2080.
[16]	YANG X， LIU Y， ZHOU S， et al. Cluster-guided contrastive graph clustering network［C］// Proceedings of the 37th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2023： 10834-10842.
[17]	HUANG Y D， ZHANG G Y， HUANG D， et al. Confidence-oriented contrastive graph clustering［C］// Proceedings of the 2024 International Joint Conference on Neural Networks. Piscataway： IEEE， 2024： 1-8.
[18]	ZHANG C Y， YAO H Y， CHEN C L P， et al. Graph representation learning via contrasting cluster assignments［J］. IEEE Transactions on Cognitive and Developmental Systems， 2024， 16（3）： 912-922.
[19]	JIN M， ZHENG Y， LI Y F， et al. Multi-scale contrastive Siamese networks for self-supervised graph representation learning［C］// Proceedings of the 30th International Joint Conference on Artificial Intelligence. California： ijcai.org， 2021： 1477-1483.
[20]	WANG T， YANG G， HE Q， et al. NCAGC： a neighborhood contrast framework for attributed graph clustering ［EB/OL］. ［2025-08-15］..
[21]	XIE X， CHEN W， KANG Z， et al. Contrastive graph clustering with adaptive filter［J］. Expert Systems with Applications， 2023， 219： No.119645.
[22]	CHEN J， KOU G. Attribute and structure preserving graph contrastive learning［C］// Proceedings of the 37th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2023： 7024-7032.
[23]	LI B， WANG Y， ZHAO B， et al. Attributed graph clustering with multi-scale weight-based pairwise coarsening and contrastive learning ［J］. Neurocomputing， 2025， 648： No.130796.
[24]	BO D， WANG X， SHI C， et al. Structural deep clustering network［C］// Proceedings of the Web Conference 2020. New York： ACM， 2020： 1400-1410.
[25]	PENG Z， LIU H， JIA Y， et al. Attention-driven graph clustering network［C］// Proceedings of the 29th ACM International Conference on Multimedia. New York： ACM， 2021： 935-943.
[26]	WANG C， PAN S， YU C P， et al. Deep neighbor-aware embedding for node clustering in attributed graphs［J］. Pattern Recognition， 2022， 122： No.108230.
[27]	GUO L， DAI Q. Graph clustering via variational graph embedding［J］. Pattern Recognition， 2022， 122： No.108334.
[28]	GUO Y， KANG L， WU M， et al. Joint node representation learning and clustering for attributed graph via graph diffusion convolution［C］// Proceedings of the 2023 International Joint Conference on Neural Networks. Piscataway： IEEE， 2023： 1-7.
[29]	TU W， GUAN R， ZHOU S， et al. Attribute-missing graph clustering network［C］// Proceedings of the 38th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2024： 15392-15401.
[30]	VELIČKOVIĆ P， FEDUS W， HAMITON W L， et al. Deep graph infomax［EB/OL］. ［2025-08-15］..

Multi-level neighborhood contrastive attribute graph clustering based on adaptive learning

基于自适应学习的多层次邻域对比属性图聚类

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 30

Related Articles 15

Recommended Articles

Metrics

[1]	Bingqing LI, Binhao HUANG, Yubei TANG, Baili ZHANG. Quality of service prediction model for data sparsity and cold start problems [J]. Journal of Computer Applications, 2026, 46(6): 1829-1835.
[2]	Hang QI, Tingting DONG, Yongqiang NAI, Xian MO. Contrastive collaborative filtering method based on graph diffusion generation and adaptive sampling [J]. Journal of Computer Applications, 2026, 46(6): 1818-1828.
[3]	Kun FU, Haoyu WEI, Weijing LIU, Xing DANG, Zezheng LIU, Jianwei LI. Graph neural network framework for topology semantic dual-domain collaboration [J]. Journal of Computer Applications, 2026, 46(5): 1378-1387.
[4]	Dirui ZHANG, Jiayu LIN, Zuhong LIANG. Supervised contrastive generative sentiment analysis method with uncertainty-aware unlikelihood learning [J]. Journal of Computer Applications, 2026, 46(5): 1416-1423.
[5]	Haihua ZHAO, Yijun HU, Rui TANG, Xian MO. Multimodal recommendation method based on semantic fusion and contrast enhancement [J]. Journal of Computer Applications, 2026, 46(4): 1058-1068.
[6]	Xiaoxia LIU, Liqun KUANG, Song WANG, Shichao JIAO, Huiyan HAN, Fengguang XIONG. Multi-scale spatio-temporal decoupling for contrastive learning of skeleton action recognition [J]. Journal of Computer Applications, 2026, 46(3): 767-774.
[7]	Yuhang XIAO, Guanfeng LI, Yuyin CHEN, Jing QIN. Few-shot relation extraction model with graph-based multi-view contrastive learning [J]. Journal of Computer Applications, 2026, 46(3): 732-740.
[8]	Hu LUO, Mingshu ZHANG. Rumor detection method based on cross-modal attention mechanism and contrastive learning [J]. Journal of Computer Applications, 2026, 46(2): 361-367.
[9]	Limei DONG, Yanzi LI, Jiayin LI, Li XU. Neighborhood-enhanced unsupervised graph anomaly detection [J]. Journal of Computer Applications, 2026, 46(2): 458-466.
[10]	Ziyang CHENG, Ruizhang HUANG, Jingjing XUE. Deep evolutionary topic clustering model [J]. Journal of Computer Applications, 2026, 46(1): 85-94.
[11]	Wen LI, Kairong LI, Kai YANG. Subgraph-aware contrastive learning with data augmentation [J]. Journal of Computer Applications, 2026, 46(1): 1-9.
[12]	Xingyao YANG, Zheng QI, Jiong YU, Zulian ZHANG, Shuai MA, Hongtao SHEN. Session-based recommendation model based on time-aware and space-enhanced dual channel graph neural network [J]. Journal of Computer Applications, 2026, 46(1): 104-112.
[13]	Chao LIU, Yanhua YU. Knowledge-aware recommendation model combining denoising strategy and multi-view contrastive learning [J]. Journal of Computer Applications, 2025, 45(9): 2827-2837.
[14]	Zhixiong XU, Bo LI, Xiaoyong BIAN, Qiren HU. Adversarial sample embedded attention U-Net for 3D medical image segmentation [J]. Journal of Computer Applications, 2025, 45(9): 3011-3016.
[15]	Zhiyuan WANG, Tao PENG, Jie YANG. Integrating internal and external data for out-of-distribution detection training and testing [J]. Journal of Computer Applications, 2025, 45(8): 2497-2506.