通过知识感知的邻居过滤机制增强社交推荐

doi:10.19665/j.issn1001-2400.20250603

摘要/Abstract

摘要：

社交推荐通过挖掘用户间的社交联系,有助于提升个性化推荐系统的性能。然而,目前大多数方法难以充分捕捉用户与物品间的复杂关系,且忽略了由于存在无共同兴趣甚至错误的社交联系所导致的社交不一致问题,从而影响用户嵌入表征的正确性和社交推荐的准确性。为此,文中提出了一种基于知识感知的邻居过滤机制用于社交推荐,旨在通过融合知识图谱与图神经网络来解决上述问题。首先,利用知识图谱嵌入技术对用户、物品及评分进行嵌入表示,从而捕捉他们之间的潜在关系模式。随后,嵌入向量被输入到图神经网络中,用于优化图神经网络节点的嵌入表示。为了提高模型的一致性识别能力,文中根据待评估的用户物品动态地构建了查询向量,并利用知识图谱综合建模查询向量与邻居节点之间的一致性分数,通过采样聚合更一致的邻居节点,从而增强图神经网络模型过滤不一致邻居节点以及节点表征的能力。在3个公开数据集上的大量实验证明了所提方法相较于现有主流方法的优越性。

关键词: 推荐系统, 图神经网络, 社交推荐, 邻居过滤

Abstract:

Social recommendation helps to improve the performance of personalized recommender systems by exploiting user social connections. However, most existing methods struggle to fully capture the complex relationships between users and items, while neglecting the issue of social inconsistency caused by irrelevant or even erroneous social ties, thus reducing the correctness of user embeddings and the accuracy of social recommendations. This paper proposes a knowledge-aware neighbor filtering mechanism for social recommendation (KFRec), aiming to resolve the aforementioned issues by integrating knowledge graphs with graph neural networks. First, this paper utilizes knowledge graph embedding techniques to vectorially represent users, items, and ratings, thereby capturing the latent relational patterns among them. Subsequently, these vectors are fed into a graph neural network to optimize the node representations of the graph neural network. To improve the model's consistency recognition capability, this paper dynamically constructs query vectors based on the user-item pairs to be evaluated, and comprehensively model the consistency scores between the query vectors and neighbor nodes using the knowledge graph. By sampling and aggregating more consistent neighbor nodes, the graph neural network model's ability to filter inconsistent neighbor nodes and node representations is enhanced. Extensive experiments on three public datasets demonstrate the superiority of our proposed KFRec over existing mainstream methods.

Key words: recommender systems, graph neural network, social recommendation, neighbor filtering

中图分类号:

TP183

周家亮, 慕彩红, 刘逸, 陈云龙. 通过知识感知的邻居过滤机制增强社交推荐[J]. 西安电子科技大学学报, 2025, 52(5): 99-110.

ZHOU Jialiang, MU Caihong, LIU Yi, CHEN Yunlong. Enhancing social recommendation via the knowledge-aware neighbor filtering mechanism[J]. Journal of Xidian University, 2025, 52(5): 99-110.

图/表 12

图1

图2

表1

表2

图3

图4

图5

图6

图7

图8

图9

表3

参考文献 31

[1]	YUN X, CHUN M H. The Impact of Personalized Recommendation on Purchase Intention under the Background of Big Data[J]. Big Data and Information Analytics, 2024, 8: 80-108.
[2]	LI X, SUN L, LING M J, et al. A Survey of Graph Neural Network Based Recommendation in Social Networks[J]. Neurocomputing, 2023, 549: 126441. doi: 10.1016/j.neucom.2023.126441
[3]	MU C H, HUANG H Y, FANG Y F, et al. A Graph Convolutional Neural Network for Recommendation Based on Community Detection and Combination of Multiple Heterogeneous Graphs[C]// Proceedings of the 2023 IEEE International Conference on Data Mining. Piscataway: IEEE, 2023: 1235-1240.
[4]	MU C H, ZHANG K Y, ZHOU J L, et al. Perturbation-Driven Dual Auxiliary Contrastive Learning for Collaborative Filtering Recommendation[C]// Proceedings of the 31st International Conference on Computational Linguistics (COLING). Piscataway: IEEE, 2025: 647-657.
[5]	黄河源, 慕彩红, 方云飞, 等. 使用图负采样的图卷积神经网络推荐算法[J]. 西安电子科技大学学报, 2024, 51(1): 86-99.
	HUANG Heyuan, MU Caihong, FANG Yunfei, et al. Graph Convolution Neural Network for Recommendation Using Graph Negative Sampling[J]. Journal of Xidian University, 2024, 51(1): 86-99.
[6]	WU L, LI J W, SUN P J, et al. DiffNet++: A Neural Influence and Interest Diffusion Network for Social Recommendation[J]. IEEE Transactions on Knowledge and Data Engineering, 2022, 34(10): 4753-4766. doi: 10.1109/TKDE.2020.3048414
[7]	MA W Z, WANG Y X, et al. MADM:A Model-Agnostic Denoising Module for Graph-Based Social Recommendation[C]// Proceedings of the 17th ACM International Conference on Web Search and Data Mining (WSDM'24). New York: ACM, 2024: 501-509.
[8]	JIANG W, GAO X Y, XU G D, et al. Challenging Low Homophily in Social Recommendation[C]// Proceedings of the ACM Web Conference 2024. New York: ACM, 2024: 3476-3484.
[9]	徐翔, 余珺君. 社交网络内容用户“茧房趋同性”——基于BERT模型的新浪微博样本研究[J]. 北京理工大学学报( 社会科学版), 2023, 25(4): 182-191.
	XU Xiang, YU Junjun. "Cocoon Homophily" of Social Network Content Users: A Study of Sina Weibo Samples Based on BERT Model[J]. Journal of Beijing Institute of Technology (Social Sciences Edition), 2023, 25(4): 182-191.
[10]	SUN Y C, SUN Z, DU Y P, et al. Self-Supervised Denoising through Independent Cascade Graph Augmentation for Robust Social Recommendation[C]// Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2024: 2806-2817.
[11]	ASHADUZZAMAN M, NGUYEN T L, TSAI C H. Explaining Social Recommendations Using Large Language Models[C]// International Conference on Disruptive Technologies, Tech Ethics and Artificial Intelligence.Berlin:Springer, 2024: 73-84.
[12]	MA H, YANG H X, LYU M R, et al. SoRec: Social Recommendation Using Probabilistic Matrix Factorization[C]// Proceedings of the 17th ACM Conference on Information and Knowledge Management. New York: ACM, 2008: 931-940.
[13]	JAMALI M, ESTER M. A Matrix Factorization Technique with Trust Propagation for Recommendation in Social Networks[C]// Proceedings of the Fourth ACM Conference on Recommender Systems. New York: ACM, 2010: 135-142.
[14]	WANG X, HE X N, NIE L Q, et al. Item Silk Road: Recommending Items from Information Domains to Social Users[C]// Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2017: 185-194.
[15]	FAN W Q, LI Q, CHENG M. Deep Modeling of Social Relations for Recommendation[C]// Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence and Thirtieth Innovative Applications of Artificial Intelligence Conference and Eighth AAAI Symposium on Educational Advances in Artificial Intelligence. New Orleans: AAAI, 2018: 1029.
[16]	荀亚玲, 李欣意, 韩硕, 等. 基于图神经网络的去偏因果推荐[J]. 计算机应用研究, 2025, 42(5): 1331-1337.
	XUN Yaling, LI Xinyi, HAN Shuo, et al. Debiased Causal Recommendation Based on Graph Neural Network[J]. Application Research of Computers, 2025, 42(5): 1331-1337.
[17]	FAN W Q, MA Y, LI Q, et al. Graph Neural Networks for Social Recommendation[C]// Proceedings of the 2019 World Wide Web Conference. New York: ACM, 2019: 417-426.
[18]	FAN W Q, MA Y, LI Q, et al. A Graph Neural Network Framework for Social Recommendations[J]. IEEE Transactions on Knowledge and Data Engineering, 2022, 34(5): 2033-2047. doi: 10.1109/TKDE.2020.3008732
[19]	YANG L W, LIU Z W, DOU Y T, et al. ConsisRec: Enhancing GNN for Social Recommendation via Consistent Neighbor Aggregation[C]// Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2021: 2141-2145.
[20]	YUAN K, LIU G N, WU J J, et al. Semantic and Structural View Fusion Modeling for Social Recommendation[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(11): 11872-11884. doi: 10.1109/TKDE.2022.3230972
[21]	WANG X, HE X N, CAO Y X, et al. KGAT: Knowledge Graph Attention Network for Recommendation[C]// Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM, 2019: 950-958.
[22]	WANG H W, ZHAO M, XIE X, et al. Knowledge Graph Convolutional Networks for Recommender Systems[C]// Proceedings of the 2019 World Wide Web Conference. New York: ACM, 2019: 3307-3313.
[23]	WANG Z, ZHANG J W, FENG J L, et al. Knowledge Graph Embedding by Translating on Hyperplanes[C]// Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence. Québec City: AAAI Press, 2014:1112-1119.
[24]	VELICKOVIC P, CUCURULL G, CASANOVA A, et al. Graph Attention Networks[C]// Proceedings of the 6th International Conference on Learning Representations. Vancouver: OpenReview.net, 2018: 1-12.
[25]	KINGMA D P, BA J. Adam: A Method for Stochastic Optimization[C]// Proceedings of the 3rd International Conference on Learning Representations. San Diego: ICLR, 2015: 1-12.
[26]	SALAKHUTDINOV R, MNIH A. Probabilistic Matrix Factorization[C]// Proceedings of the 21st International Conference on Neural Information Processing Systems. New York: Curran Associates Inc., 2007: 1257-1264.
[27]	VAN DEN BERG R, KIPF T N, WELLING M. Graph Convolutional Matrix Completion[C]// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM, 2018: 165-174.
[28]	CHEN J J, XIN X, LIANG X F, et al. GDSRec: Graph-Based Decentralized Collaborative Filtering for Social Recommendation[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(5): 4813-4824.
[29]	GLOROT X, BENGIO Y. Understanding the Difficulty of Training Deep Feedforward Neural Networks[C]// Proceedings of the International Conference on Artificial Intelligence and Statistics. Piscataway: IEEE, 2010:249-256.
[30]	VAN DER MAATEN L, HINTON G. Visualizing Data Using t-SNE[J]. Journal of Machine Learning Research, 2008, 9(2605): 2579-2605.
[31]	YU J L, YIN H Z, XIA X, et al. Are Graph Augmentations Necessary? Simple Graph Contrastive Learning for Recommendation[C]// Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2022: 1294-1303.

统计指标	Ciao	Epinions	Yelp
用户数量	7 317	18 069	21 374
物品数量	104 975	261 246	109 420
评分数量	283 319	762 938	920 354
评分密度/%	0.036 8	0.016 1	0.039 3
社交联系	111 781	355 530	497 206
社交密度/%	0.208 7	0.108 9	0.108 8

模型	Ciao		Epinions		Yelp
模型	RMSE	MAE	RMSE	MAE	RMSE	MAE
PMF	1.123 8	0.902 1	1.212 8	0.995 2	1.546 7	1.009 2
SoRec	1.065 2	0.841 0	1.143 7	0.896 1	1.370 1	0.969 7
SocialMF	1.050 1	0.827 0	1.132 8	0.883 7	1.305 8	0.954 1
GCMC+SN	0.993 1	0.752 6	1.071 1	0.859 0	1.145 9	0.887 1
GraphRec	0.987 8	0.744 7	1.060 4	0.823 8	1.044 4	0.809 5
ConsisRec	0.970 1	0.724 5	1.057 2	0.811 2	1.047 8	0.810 8
GraphRec+	0.968 7	0.723 3	1.057 9	0.809 9	1.042 1	0.808 9
GDSRec	0.967 0	0.727 3	1.056 5	0.808 9	1.049 9	0.810 2
KFRec	0.954 9	0.708 1	1.050 1	0.807 6	1.040 5	0.804 9

模型	Ciao	Epinions
GraphRec	16.42	48.80
ConsisRec	13.30	47.50
GDSRec	12.05	28.51
KFRec	20.43	53.41