Multi-Relationship Recommendation Model Based on User Interest-Aware

doi:10.3778/j.issn.1002-8331.2206-0148

Abstract

Abstract: Because graph convolution network（GCN） can use the cooperation signals of high-order neighbors to learn the embedding of users and items better, it has been widely used in recommendation systems. However, in the current multi-relationship recommendation model based on GCN, the embedding learning of user nodes will be interfered by high-order adjacent users with dissimilar interests, so that users with different interests will get similar embedding after multi-layer graph convolution, resulting in the over-smoothing problem. Therefore, this paper proposes a multi-relationship recommendation model based on user interest-aware（IMRRM） because the above problem. First, the model uses a light graph convolutional network in the user-item heterogeneous interaction graph to obtain the graph structure information of each user. And then the subgraph generation module uses the user’s graph structure information and initial features to effectively identify users with similar interests, and group similar users and their interaction items into a subgraph. Finally, more accurate user embeddings are obtained by performing deep embedding learning in subgraphs to prevent unrelated high-order neighbors from propagating more negative information. Therefore, the IMRRM model reduces the influence of noise information on user embedding learning, effectively alleviates the over-smoothing problem, and makes multi-relationship recommendation more accurately. The effectiveness and robustness of IMRRM are verified by experiments on two public datasets, Beibei and Taobao. The experimental results show that the IMRRM model is improved by 1.98% and 1.49% on HR10, and 1.58% and 1.81% on NDCG10, respectively, with better performance.

Key words: graph convolutional network, multi-relational recommendation, subgraph, interest-aware

摘要： 由于图卷积网络（GCN）能够利用高阶邻居的协作信号来更好地学习用户和项目的嵌入，它已经广泛应用于推荐系统。但在当前基于GCN的多关系推荐模型中用户节点的嵌入学习会受到与之兴趣不相似的高阶相邻用户的干扰，导致拥有不同兴趣的用户经过多层图卷积后会得到相似的嵌入，从而产生了过度平滑问题。因此针对上述问题提出了一个基于用户兴趣感知的多关系推荐模型（IMRRM）。该模型会在用户项目异构交互图中利用轻量化的图卷积网络得到每个用户的图形结构信息。子图生成模块利用用户的图结构信息和初始特征有效地识别出兴趣相似的用户，并将相似用户及其交互项目组成一个子图。通过在子图中进行深层嵌入学习来防止兴趣不相关的高阶邻居传播更多的负面信息从而得到更精确的用户嵌入。因此IMRRM模型减少了噪声信息对用户节点嵌入学习的影响，有效地缓解了过度平滑问题来更加准确地进行多关系推荐。通过在Beibei和Taobao这两个公共数据集上实验来验证IMRRM的有效性和鲁棒性。实验结果表明，IMRRM模型在HR10上分别提高了1.98%和1.49%，在NDCG10上分别提高了1.58%和1.81%，具有较好的性能。

关键词: 图卷积网络, 多关系推荐, 子图, 兴趣感知

HU Xinrong, DENG Jiewen, LUO Ruiqi, LIU Junping, ZHU Qiang, PENG Tao. Multi-Relationship Recommendation Model Based on User Interest-Aware[J]. Computer Engineering and Applications, 2023, 59(11): 231-240.

胡新荣, 邓杰文, 罗瑞奇, 刘军平, 朱强, 彭涛. 基于用户兴趣感知的多关系推荐模型[J]. 计算机工程与应用, 2023, 59(11): 231-240.

References

[1] RICCI F，ROKACH L，SHAPIRA B.Introduction to recommender systems handbook[M]//Recommender systems handbook.Berlin：Springer，2011：1-35.
[2] EBESU T，SHEN B，FANG Y.Collaborative memory network for recommendation systems[C]//Proceedings of SIGIR，2018：515-524.
[3] HE X，LIAO L，ZHANG H，et al.Neural collaborative filtering[C]//The 26th International Conference on World Wide Web，2017：173-182.
[4] LIANG D，KRISHNAN R G，HOFFMAN M D，et al.Variational autoencoders for collaborative filtering[C]//The World Wide Web Conference，2018：689-698.
[5] DING J，YU G，HE X，et al.Improving implicit recommender systems with view data[C]//Proceedings of IJCAI，2018：3343-3349.
[6] GAO C，HE X，GAN D，et al.Neural multi-task recommendation from multi-behavior data[C]//The 35th International Conference on Data Engineering（ICDE），2019.
[7] CHEN C，ZHANG M，ZHANG Y，et al.Efficient heterogeneous collaborative filtering without negative sampling for recommendation[C]//Proceedings of the AAAI Conference on Artificial Intelligence，2020：19-26.
[8] XIA L，XU Y，HUANG C，et al.Graph meta network for multi-behavior recommendation[C]//The 44th International ACM SIGIR Conference on Research and Develop-ment in Information Retrieval，2021.
[9] 张程东，王绍卿，刘玉芳，等.采用新型元路径的异构图表示学习方法[J/OL].计算机科学与探索：1-11[2022-07-25].http：//kns.cnki.net/kcms/detail/11.5602.TP.20220610.1030.
006.html.
ZHANG C D，WANG S Q，LIU Y F，et al.A method of heterogeneous graph representation learning using novel meta-path[J/OL].Journal of Frontiers of Computer Science and Technology：1-11[2022-07-25].http：//kns.cnki.net/kcms/detail/11/5602.TP.20220610.1030.006.html.
[10] JIN B，GAO C，HE X，et al.Multi-behavior recommendation with graph convolutional networks[C]//The 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval，2020.
[11] CHEN C，MA W，ZHANG M，et al.Graph heterogeneous multi-relational recommendation[C]//Proceedings of AAAI，2021.
[12] LIU F，CHENG Z，ZHU L，et al.Interest-aware message-passing GCN for recommendation[C]//Proceedings of the Web Conference，2021.
[13] WU F，ZHANG T，SOUZA A，et al.Simplifying graph convolutional networks[C]//Proceedings of the 36th International Conference on Machine Learning，Long Beach，California，USA，2019：6861-6871.
[14] LIU Z，DOU Y，YU P S，et al.Alleviating the inconsistency problem of applying graph neural network to fraud detection[C]//Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval，2020：1569-1572.
[15] BRUNA J，ZAREMBA W，SZLAM A，et al.Spectral networks and locally connected networks on graphs[J].arXiv：1312.6203，2013.
[16] HAMILTON W L，YING Z，LESKOVEC J.Inductive representation learning on large graphs[C]//Proceedings of NIPS，2017：1024-1034.
[17] VELICKOVIC P，CUCURULL G，CASANOVA A，et al.Graph attention networks[J].arXiv：1710.10903，2017.
[18] MARCHEGGIANI D，TITOV I.Encoding sentences with graph convolutional networks for semantic role labeling[C]//Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing，2017：1506-1515.
[19] SCHLICHTKRULL M S，KIPF T，BLOEM P，et al.Modeling relational data with graph convolutional networks[M].Cham：Springer，2018：593-607.
[20] BERG R，KIPF T N，WELLING M.Graph convolutional matrix completion[J].arXiv：1706.02263，2017.
[21] YING R，HE R，CHEN K，et al.Graph convolutional neural networks for webscale recommender systems[C]//Proceedings of SIGKDD，2018：974-983.
[22] ZHENG L，LU C，JIANG F，et al.Spectral collaborative filtering[C]//Proceedings of RecSys，2018：311-319.
[23] WANG X，HE X，WANG M，et al.Neural graph collabo-rative filtering[C]//Proceedings of SIGIR，2019：165-174.
[24] CHEN L，WU L，HONG R，et al.Revisiting graph based collaborative filtering：a linear residual graph convolutional network approach[C]//Proceedings of the AAAI Conference on Artificial Intelligence，2020：27-34.
[25] HE X，DENG K，WANG X，et al.LightGCN：simplifying and powering graph convolution network for recommendation[J].arXiv：2002.02126，2020.
[26] XUE F，HE X，WANG X，et al.Deep item-based collabo-rative filtering for top-n recommendation[J].ACM Transactions on Information Systems，2019，37：1-25.
[27] KIP F T N，WELLING M.Semi-supervised classification with graph convolutional networks[C]//Proceedings of ICLR，2017.
[28] HU Y，REN P，LUO W，et al.Multi-resolution representation with recurrent neural networks application for streaming time series in IoT[J].Computer Networks，2019，152：114-132.
[29] MAAS A L，HANNUN A Y，NG A Y.Rectifier nonlinearities improve neural network acoustic models[C]//Proceedings of ICML，2013.
[30] 翟正利，冯舒，李鹏辉.改进图卷积网络的协同过滤推荐模型[J].计算机工程与应用，2022，58（17）：199-205.
ZHAI Z L，FENG S，LI P H.Collaborative filtering recommendation model based on improved graph convolutional network[J].Computer Engineering and Applications，2022，58（17）：199-205.
[31] CHEN C，ZHANG M，ZHANG Y，et al.Efficient neural matrix factorization without sampling for recommendation[J].ACM Transactions on Information Systems，2020，38：1-28.