融合知识图谱与高阶信息聚合机制的推荐模型

doi:10.3778/j.issn.1002-8331.2407-0354

摘要/Abstract

摘要： 知识图谱通过其丰富的语义信息和复杂的关系网络，显著提升了推荐模型的精度和可解释性。然而，现有基于嵌入传播的知识图谱推荐模型在聚合高阶信息时，难以在捕捉多层次语义关联和抑制多层传播引入的噪声之间实现有效权衡；同时，在信息传播过程中，高连接度节点往往主导特征更新，导致低连接度节点的个性化特征被稀释，削弱了模型对细粒度个性化需求的捕捉能力。针对这些问题，提出了一种融合知识图谱与高阶信息聚合机制的推荐模型。该模型以实体的邻域信息为感知域，通过多次迭代传播，有效捕捉知识图谱中的高阶连通性和复杂关系。引入对称归一化机制，以解决节点聚合过程中因度分布不均导致的特征更新偏差，确保不同实体在嵌入空间中的均衡表示。设计高阶聚合传播机制，动态整合不同层次的邻域特征信息，兼顾高阶语义信息的获取与多层传播所引入的噪声抑制。该模型在Last-FM和Book-Crossing公共数据集上与基线模型进行对比实验，结果表明该模型在AUC、F1、Recall@k和NDCG@k评价指标上优于其他模型。

关键词: 推荐模型, 知识图谱, 图卷积神经网络

Abstract: The knowledge graph significantly enhances the accuracy and interpretability of recommendation models through its rich semantic information and complex relational network. Existing knowledge graph-based recommendation models that utilize embedding propagation face challenges in balancing the capture of multi-level semantic associations and the suppression of noise introduced by multi-hop propagation during the aggregation of high-order information. Additionally, during the information propagation process, high-degree nodes tend to dominate feature updates, which dilutes the personalized features of low-degree nodes, thereby weakening the model’s ability to capture fine-grained personali-
zation. To address these issues, this paper proposes a recommendation model that integrates knowledge graph and high-order information aggregation mechanisms. Firstly, the model uses the neighborhood information of entities as the perceptual domain and, through multiple iterations of propagation, effectively captures the high-order connectivity and complex relations within the knowledge graph. Secondly, a symmetric normalization mechanism is introduced to address the feature update bias caused by degree distribution imbalance during node aggregation, ensuring balanced representations of different entities in the embedding space. Finally, a high-order aggregation propagation mechanism is designed to dynamically integrate multi-level neighborhood features, balancing the capture of high-order semantic information with noise suppression introduced by multi-hop propagation. Comparative experiments with baseline models on the Last-FM and Book-Crossing public datasets demonstrate that the model outperforms other advanced models in AUC, F1, Recall@k and NDCG@k indicators.

Key words: recommendation model, knowledge graph, graph convolutional neural network

武杲昊, 王霞, 郝国生, 祝义. 融合知识图谱与高阶信息聚合机制的推荐模型[J]. 计算机工程与应用, 2025, 61(19): 158-166.

WU Gaohao, WANG Xia, HAO Guosheng, ZHU Yi. Recommendation Model Integrating Knowledge Graph and High-Order Information Aggregation Mechanism[J]. Computer Engineering and Applications, 2025, 61(19): 158-166.

参考文献

[1] 张明星, 张骁雄, 刘姗姗, 等. 利用知识图谱的推荐系统研究综述[J]. 计算机工程与应用, 2023, 59(4): 30-42.
ZHANG M X, ZHANG X X, LIU S S, et al. Review of recommendation systems using knowledge graph[J]. Computer Engineering and Applications, 2023, 59(4): 30-42.
[2] ZHAI F B, LI B Z. Point of interest recommendation based on graph convolutional neural network[J]. Journal of Physics: Conference Series, 2021, 1883(1): 012132.
[3] JI S X, PAN S R, CAMBRIA E, et al. A survey on knowledge graphs: representation, acquisition, and applications[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022, 33(2): 494-514.
[4] GUO Q Y, ZHUANG F Z, QIN C, et al. A survey on knowledge graph-based recommender systems[J]. IEEE Transactions on Knowledge and Data Engineering, 2022, 34(8): 3549-3568.
[5] WANG H W, ZHANG F Z, XIE X, et al. DKN: deep knowledge-aware network for news recommendation[C]//Proceedings of the 2018 World Wide Web Conference. New York: ACM, 2018: 1835-1844.
[6] WANG H W, ZHANG F Z, ZHAO M, et al. Multi-task feature learning for knowledge graph enhanced recommendation[C]//Proceedings of the World Wide Web Conference. New York: ACM, 2019: 2000-2010.
[7] HUANG X W, FANG Q, QIAN S S, et al. Explainable interaction-driven user modeling over knowledge graph for sequential recommendation[C]//Proceedings of the 27th ACM International Conference on Multimedia. New York: ACM, 2019: 548-556.
[8] MA W Z, ZHANG M, CAO Y, et al. Jointly learning explainable rules for recommendation with knowledge graph[C]//Proceedings of the World Wide Web Conference. New York: ACM, 2019: 1210-1221.
[9] TU K, CUI P, WANG D X, et al. Conditional graph attention networks for distilling and refining knowledge graphs in recommendation[C]//Proceedings of the 30th ACM International Conference on Information & Knowledge Management. New York: ACM, 2021: 1834-1843.
[10] 张彬, 郝利新, 张国防. 融合元图邻域的知识图谱推荐模型[J]. 计算机应用研究, 2024, 41(8): 2412-2418.
ZHANG B, HAO L X, ZHANG G F. Knowledge graph recommendation model with integrated meta-graph neighborhoods[J]. Application Research of Computers, 2024, 41(8): 2412-2418.
[11] LIN Y K, LIU Z Y, SUN M S, et al. Learning entity and relation embeddings for knowledge graph completion[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2015.
[12] WANG Z, ZHANG J W, FENG J L, et al. Knowledge graph embedding by translating on hyperplanes[C]//Proceedings of the 28th AAAI Conference on Artificial Intelligence, 2014.
[13] NICKEL M, TRESP V, KRIEGAL H P. A three-way model for collective learning on multi-relational data[C]//Proceedings of the 28th International Conference on Machine Learning, 2011: 809-816.
[14] YANG B, YIH W, HE X, et al. Embedding entities and relations for learning and inference in knowledge bases[J]. arXiv:1412.6575, 2014.
[15] CAO Y X, WANG X, HE X N, et al. Unifying knowledge graph learning and recommendation: towards a better understanding of user preferences[C]//Proceedings of the World Wide Web Conference. New York: ACM, 2019: 151-161.
[16] ZHANG C Y, WANG Y, ZHU L, et al. Multi-graph heterogeneous interaction fusion for social recommendation[J]. ACM Transactions on Information Systems, 2022, 40(2): 1-26.
[17] WANG X, HUANG T L, WANG D X, et al. Learning intents behind interactions with knowledge graph for recommendation[C]//Proceedings of the Web Conference 2021. New York: ACM, 2021: 878-887.
[18] 张小婉, 邓秋军, 柳先辉. 结合图注意力机制的知识图谱推荐算法[J]. 计算机科学, 2023, 50(S2): 464-470.
ZHANG X W, DENG Q J, LIU X H. Knowledge map recommendation algorithm based on graph attention mechanism[J]. Computer Science, 2023, 50(S2): 464-470.
[19] 朱冬亮, 文奕, 万子琛. 基于知识图谱的推荐系统研究综述[J]. 数据分析与知识发现, 2021, 5(12): 1-13.
ZHU D L, WEN Y, WAN Z C. Review of recommendation systems based on knowledge graph[J]. Data Analysis and Knowledge Discovery, 2021, 5(12): 1-13.
[20] YANG Y H, HUANG C, XIA L H, et al. Knowledge graph self-supervised rationalization for recommendation[C]//Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2023: 3046-3056.
[21] LYU Z Y, WU Y, LAI J J, et al. Knowledge enhanced graph neural networks for explainable recommendation[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(5): 4954-4968.
[22] XU Z M, LIU H L, LI J, et al. CKGAT: collaborative knowledge-aware graph attention network for top-N recommendation[J]. Applied Sciences, 2022, 12(3): 1669.
[23] CHEN Y K, YANG M L, ZHANG Y X, et al. Modeling scale-free graphs with hyperbolic geometry for knowledge-aware recommendation[C]//Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining. New York: ACM, 2022: 94-102.
[24] ZHAO N, LONG Z, WANG J, et al. AGRE: a knowledge graph recommendation algorithm based on multiple paths embeddings RNN encoder[J]. Knowledge-Based Systems, 2023, 259: 110078.
[25] WANG H, ZHANG F, ZHAO M, et al. Knowledge graph convolutional networks for recommender systems[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2019: 973-982.
[26] LIU Y, YANG S S, XU Y H, et al. Contextualized graph attention network for recommendation with item knowledge graph[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(1): 181-195.
[27] MA T, HUANG L T, LU Q Q, et al. KR-GCN: knowledge-aware reasoning with graph convolution network for explainable recommendation[J]. ACM Transactions on Information Systems, 2023, 41(1): 1-27.