融合知识图谱与图卷积网络的混合推荐模型

doi:10.3778/j.issn.1673-9418.2110057

计算机科学与探索 ›› 2022, Vol. 16 ›› Issue (6): 1343-1353.DOI: 10.3778/j.issn.1673-9418.2110057

融合知识图谱与图卷积网络的混合推荐模型

郭晓旺¹, 夏鸿斌¹^,²^,⁺(), 刘渊¹^,²

1. 江南大学人工智能与计算机学院,江苏无锡 214122
2. 江苏省媒体设计与软件技术重点实验室,江苏无锡 214122

收稿日期:2021-10-22 修回日期:2022-01-25 出版日期:2022-06-01 发布日期:2022-06-20
通讯作者: + E-mail: hbxia@163.com
作者简介:郭晓旺（1996—）,女,河南安阳人,硕士研究生,主要研究方向为机器学习、推荐系统。
夏鸿斌（1972—）,男,安徽庐江人,博士,副教授,CCF会员,主要研究方向为个性化推荐、自然语言处理、网络优化。
刘渊（1967—）,男,江苏无锡人,教授,CCF高级会员,主要研究方向为网络安全、社交网络。
基金资助:
国家自然科学基金(61972182)

Hybrid Recommendation Model of Knowledge Graph and Graph Convolutional Network

GUO Xiaowang¹, XIA Hongbin¹^,²^,⁺(), LIU Yuan¹^,²

1. School of Artificial Intelligence and Computer, Jiangnan University, Wuxi, Jiangsu 214122, China
2. Jiangsu Key Laboratory of Media Design and Software Technology, Wuxi, Jiangsu 214122, China

Received:2021-10-22 Revised:2022-01-25 Online:2022-06-01 Published:2022-06-20
About author:GUO Xiaowang, born in 1996, M.S. candidate.Her research interests include machine learning and recommendation system.
XIA Hongbin, born in 1972, Ph.D., associate professor, member of CCF. His research interests include personalized recommendation, natural language processing and network optimization.
LIU Yuan, born in 1967, professor, senior mem-ber of CCF. His research interests include network security and social network.
Supported by:
National Natural Science Foundation of China(61972182)

摘要/Abstract

摘要：

针对当前多数基于知识图谱的推荐模型未能充分对用户特征建模,且未考虑知识图谱中实体间的邻域关系的问题,提出了一种融合知识图谱与图卷积网络的混合推荐模型（HKC）。首先,利用KGCN算法捕捉项目间的相关性,通过邻域聚合计算得到项目的特征向量;然后,通过协作传播提取知识图谱中与用户相联系的实体,使用交替学习的方式同时优化模型预测单元和知识图谱嵌入单元,通过交互单元计算得到用户的特征向量;最后,将用户特征向量和项目特征向量送入预测环节,通过向量的内积运算以及归一化操作计算用户与项目的交互概率。在三种公开数据集上与七个基线模型进行了对比实验,在MovieLens-1M数据集上,AUC提升了0.25%~37.41%,ACC提升了0.78%~49.44%;在Book-Crossing数据集上,AUC提升了0.04%~19.38%,ACC提升了6.49%~18.60%;在Last.FM数据集上,AUC提升了1.33%~33.50%,ACC提升了0.36%~30.66%。实验结果表明,提出的混合推荐模型与其他具有代表性的推荐模型相比具有良好的推荐性能。

关键词: 推荐系统, 知识图谱, 交替学习, 邻域聚合

Abstract:

Concerning the failure of most current recommendation models based on knowledge graph to adequately model users’ characteristics, the neighborhood relationship between entities in the knowledge graph is not considered. This paper proposes a hybrid recommendation model that combines knowledge graph and graph convolutional network (HKC). Firstly, the KGCN (knowledge graph convolutional networks for recommender systems) algorithm is used to capture the correlation between items, and obtain the feature vector of the item through neighborhood aggregation unit. The entities associated with the user in the knowledge graph are extracted through collaborative propagation. Then the model uses the alternate learning method to optimize the model prediction unit and the knowledge graph embedding unit at the same time, and calculate the user’s feature vector through the interaction unit. Finally, the user feature vector and the item feature vector are sent to the prediction link and the interaction probability between the user and the item is calculated through the inner product operation and normalization of the vector. Comparative experiments are conducted on three public datasets with seven baseline models. On the MovieLens-1M dataset, AUC is increased by 0.25% to 37.41%, and ACC is increased by 0.78% to 49.44%; on the Book-Crossing dataset, AUC is increased by 0.04% to 19.38%, and ACC is increased by 6.49% to 18.60%; on the Last.FM dataset, AUC is increased by 1.33% to 33.50%, and ACC is increased by 0.36% to 30.66%. Experimental results show that the model proposed has improved performance compared with other benchmark models.

Key words: recommender system, knowledge graph, alternate learning, neighborhood aggregation

中图分类号:

TP391

郭晓旺, 夏鸿斌, 刘渊. 融合知识图谱与图卷积网络的混合推荐模型[J]. 计算机科学与探索, 2022, 16(6): 1343-1353.

GUO Xiaowang, XIA Hongbin, LIU Yuan. Hybrid Recommendation Model of Knowledge Graph and Graph Convolutional Network[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(6): 1343-1353.

图/表 11

图1 HKC模型结构

Fig.1 Framework of HKC

图2 邻域聚合

Fig.2 Neighborhood aggregation

表1 三个数据集的基本统计数据

Table 1 Basic statistics of three datasets

Category	MovieLens-1M	Book-Crossing	Last.FM
Users	6 036	17 860	1 872
Items	2 347	14 910	3 846
Interaction	753 772	139 746	42 346
Entities	6 279	24 039	9 366
Relations	7	10	60
KGtriples	20 195	19 793	15 518

表2 HKC模型的超参数设置

Table 2 Hyper-parameter settings of HKC

Dataset	$d$	$K$	$H$	$t$	$λ 1$	$λ 2$
MovieLens-1M	8	8	3	2	0.5	1E-6
Book-Crossing	8	8	2	3	0.1	5E-7
Last.FM	4	4	2	2	0.1	1E-6

表2 HKC模型的超参数设置

Table 2 Hyper-parameter settings of HKC

Dataset	$d$	$K$	$H$	$t$	$λ 1$	$λ 2$
MovieLens-1M	8	8	3	2	0.5	1E-6
Book-Crossing	8	8	2	3	0.1	5E-7
Last.FM	4	4	2	2	0.1	1E-6

表3 CTR预测中的AUC和ACC结果

Table 3 Results of AUC and ACC in CTR prediction

Model	MovieLens-1M		Book-Crossing		Last.FM
Model	AUC	ACC	AUC	ACC	AUC	ACC
LibFM	0.892 3	0.812 0	0.684 7	0.640 0	0.778 1	0.708 9
DKN	0.671 2	0.568 0	0.620 7	0.588 3	0.603 2	0.573 1
KGNN_LS	0.913 7	0.840 5	0.689 7	0.635 3	0.795 4	0.726 0
$R i p p l e N e t$	0.920 0	0.842 2	0.720 4	0.647 7	0.789 1	0.722 0
KGCN	0.906 7	0.831 3	0.694 4	0.635 4	0.794 1	0.723 6
MKR	0.915 4	0.842 0	0.734 2	0.702 2	0.794 7	0.746 1
$C K A N$	0.915 2	0.840 5	0.740 7	0.655 2	0.841 0	0.744 4
HKC	0.922 3	0.848 8	0.741 0	0.697 7	0.805 3	0.748 0

表3 CTR预测中的AUC和ACC结果

Table 3 Results of AUC and ACC in CTR prediction

Model	MovieLens-1M		Book-Crossing		Last.FM
Model	AUC	ACC	AUC	ACC	AUC	ACC
LibFM	0.892 3	0.812 0	0.684 7	0.640 0	0.778 1	0.708 9
DKN	0.671 2	0.568 0	0.620 7	0.588 3	0.603 2	0.573 1
KGNN_LS	0.913 7	0.840 5	0.689 7	0.635 3	0.795 4	0.726 0
$R i p p l e N e t$	0.920 0	0.842 2	0.720 4	0.647 7	0.789 1	0.722 0
KGCN	0.906 7	0.831 3	0.694 4	0.635 4	0.794 1	0.723 6
MKR	0.915 4	0.842 0	0.734 2	0.702 2	0.794 7	0.746 1
$C K A N$	0.915 2	0.840 5	0.740 7	0.655 2	0.841 0	0.744 4
HKC	0.922 3	0.848 8	0.741 0	0.697 7	0.805 3	0.748 0

表4 模型消融研究

Table 4 Model ablation study

Model	MovieLens-1M		Book-Crossing		Last.FM
Model	AUC	ACC	AUC	ACC	AUC	ACC
HKC	0.922 3	0.848 8	0.741 0	0.697 7	0.805 3	0.748 0
HKC-n	0.921 0	0.845 5	0.695 4	0.611 8	0.778 0	0.713 1
HKC-c	0.920 1	0.846 0	0.746 3	0.702 0	0.798 0	0.747 1

图3 三种数据集上Top-K推荐的精确率

Fig.3 Precision@K in Top-K recommendation on three datasets

图4 三种数据集上Top-K推荐的召回率

Fig.4 Recall@K in Top-K recommendation on three datasets

表5 不同优化器对HKC模型性能的研究

Table 5 Research on performance of HKC model by different optimizers

Optimizer	MovieLens-1M
Optimizer	AUC	ACC
SGD	0.922 3	0.847 7
RMSprop	0.918 9	0.843 5
Adamax	0.920 5	0.846 1
Adam	0.922 3	0.848 8

图5 HKC模型在MovieLens-1M上的参数灵敏度

Fig.5 Parameter sensitivity of HKC on MovieLens-1M

图6 MovieLens-1M数据集上稀疏性场景下模型的性能

Fig.6 Model performance in sparse scenarios on MovieLens-1M

参考文献 24

[1]	KOREN Y, BELL R M, VOLINSKY C. Matrix factoriza-tion techniques for recommender systems[J]. IEEE Computer, 2009, 42(8): 30-37.
[2]	JAMALI M, ESTER M. TrustWalker: a random walk model for combining trust-based and item-based recommendation[C]// Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Paris, Jun 28-Jul 1, 2009. New York: ACM, 2009: 397-406.
[3]	WANG Y Q, SHANG W Q. Personalized news recommen-dation based on consumers’ click behavior[C]// Proceedings of the 12th International Conference on Fuzzy Systems and Knowledge Discovery, Zhangjiajie, Aug 15-17, 2015. Pisca-taway: IEEE, 2015: 634-638.
[4]	WANG J, WANG H W, ZHAO M, et al. Joint topic-semantic-aware social matrix factorization for online voting recom-mendation[J]. Knowledge-Based Systems, 2020, 210: 106433. DOI URL
[5]	WANG H W, ZHANG F Z, HOU M, et al. SHINE: signed heterogeneous information network embedding for sen-timent link prediction[C]// Proceedings of the 11th ACM International Conference on Web Search and Data Mining, Los Angeles, Feb 5-9, 2018. New York: ACM, 2018: 592-600.
[6]	ZHANG D H, LIU L N, WEI Q, et al. Neighborhood agg-regation collaborative filtering based on knowledge graph[J]. Applied Sciences, 2020, 10(11): 3818. DOI URL
[7]	ZHANG F Z, YUAN J, LIAN D F, et al. Collaborative know-ledge base embedding for recommender systems[C]// Pro-ceedings of the 22nd ACM SIGKDD International Confe-rence on Knowledge Discovery and Data Mining, San Fran-cisco, Aug 13-17, 2016. New York: ACM, 2016: 353-362.
[8]	ZHANG Y F, AI Q Y, CHEN X, et al. Learning over knowledge-base embeddings for recommendation[J]. arXiv: 1803. 06540, 2018.
[9]	WANG Q, MAO Z D, WANG B, et al. Knowledge graph embedding: a survey of approaches and applications[J]. IEEE Transactions on Knowledge and Data Engineering, 2017, 29(12): 2724-2743. DOI URL
[10]	高仰, 刘渊. 融合知识图谱和短期偏好的推荐算法[J]. 计算机科学与探索, 2021, 15(6): 1133-1144.
	GAO Y, LIU Y. Recommendation algorithm combining know-ledge graph and short-term preferences[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(6): 1133-1144.
[11]	WANG H W, ZHANG F Z, ZHAO M, et al. Multi-task fea-ture learning for knowledge graph enhanced recommen-dation[C]// Proceedings of the 2019 World Wide Web Con-ference, San Francisco, May 13-17, 2019. New York: ACM, 2019: 2000-2010.
[12]	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, San Francisco, May 13-17, 2019. New York: ACM, 2019: 3307-3313.
[13]	WANG H W, ZHANG F Z, WANG J L, et al. RippleNet: propagating user preferences on the knowledge graph for recommender systems[C]// Proceedings of the 27th ACM International Conference on Information and Knowledge Management, Torino, Oct 22-26, 2018. New York: ACM, 2018: 417-426.
[14]	ZHOU J, CUI G, ZHANG Z, et al. Graph neural networks: a review of methods and applications[J]. AI Open, 2020, 1: 57-81. DOI URL
[15]	YING R, HE R N, CHEN K F, et al. Graph convolutional neural networks for web-scale recommender systems[C]// Proceedings of the 24th ACM SIGKDD International Con-ference on Knowledge Discovery and Data Mining, London, Aug 19-23, 2018. New York: ACM, 2018: 974-983.
[16]	YANG Z X, DONG S B. HAGERec: hierarchical attention graph convolutional network incorporating knowledge graph for explainable recommendation[J]. Knowledge-Based Sys-tems, 2020, 204: 106194.
[17]	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, Lyon, Apr 23-27, 2018. New York: ACM, 2018: 1835-1844.
[18]	YU X, REN X, SUN Y Z, et al. Personalized entity recom-mendation: a heterogeneous information network approach[C]// Proceedings of the 7th ACM International Conference on Web Search and Data Mining, New York, Feb 24-28, 2014. New York: ACM, 2014: 283-292.
[19]	SUN Z, YANG J, ZHANG J, et al. Recurrent knowledge graph embedding for effective recommendation[C]// Procee-dings of the 12th ACM Conference on Recommender Systems, Vancouver, Oct 2-7, 2018. New York: ACM, 2018: 297-305.
[20]	WANG X, WANG D X, XU C R, et al. Explainable reasoning over knowledge graphs for recommendation[C]// Procee-dings of the 33rd AAAI Conference on Artificial Intelli-gence, the 31st Innovative Applications of Artificial Intel-ligence Conference, the 9th AAAI Symposium on Educa-tional Advances in Artificial Intelligence, Honolulu, Jan 27-Feb 1, 2019. Menlo Park: AAAI, 2019: 5329-5336.
[21]	李世宝, 张益维, 刘建航, 等. 基于知识图谱共同邻居排序采样的推荐模型[J]. 电子与信息学报, 2021, 43(12): 3522-3529.
	LI S B, ZHANG Y W, LIU J H, et al. Recommendation model based on public neighbor sorting and sampling of knowledge graph[J]. Journal of Electronics and Information Technology, 2021, 43(12): 3522-3529.
[22]	RENDLE S. Factorization machines with LibFM[J]. ACM Transactions on Intelligent Systems and Technology, 2012, 3(3): 57.
[23]	WANG H W, ZHANG F Z, ZHANG M D, et al. Knowledge-aware graph neural networks with label smoothness regu-larization for recommender systems[C]// Proceedings of the 25th ACM SIGKDD International Conference on Know-ledge Discovery and Data Mining, Anchorage, Aug 4-8, 2019. New York: ACM, 2019: 968-977.
[24]	WANG Z, LIN G Y, TAN H B, et al. CKAN: collaborative knowledge-aware attentive network for recommender sys-tems[C]// Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2020: 219-228.

融合知识图谱与图卷积网络的混合推荐模型

Hybrid Recommendation Model of Knowledge Graph and Graph Convolutional Network

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 24

相关文章 15

编辑推荐

Metrics

[1]	于慧琳, 陈炜, 王琪, 高建伟, 万怀宇. 使用子图推理实现知识图谱关系预测[J]. 计算机科学与探索, 2022, 16(8): 1800-1808.
[2]	萨日娜, 李艳玲, 林民. 知识图谱推理问答研究综述[J]. 计算机科学与探索, 2022, 16(8): 1727-1741.
[3]	田萱, 陈杭雪. 推荐任务中知识图谱嵌入应用研究综述[J]. 计算机科学与探索, 2022, 16(8): 1681-1705.
[4]	韩毅, 乔林波, 李东升, 廖湘科. 知识增强型预训练语言模型综述[J]. 计算机科学与探索, 2022, 16(7): 1439-1461.
[5]	陈江美, 张文德. 基于位置社交网络的兴趣点推荐系统研究综述[J]. 计算机科学与探索, 2022, 16(7): 1462-1478.
[6]	董文波, 孙仕亮, 殷敏智. 医学知识推理研究现状与发展[J]. 计算机科学与探索, 2022, 16(6): 1193-1213.
[7]	王宝亮, 潘文采. 基于知识图谱的双端邻居信息融合推荐算法[J]. 计算机科学与探索, 2022, 16(6): 1354-1361.
[8]	武森, 董雅贤, 魏桂英, 高晓楠. 面向稀疏数据的协同过滤用户相似度计算研究[J]. 计算机科学与探索, 2022, 16(5): 1043-1052.
[9]	张子辰, 岳昆, 祁志卫, 段亮. 时序知识图谱的增量构建[J]. 计算机科学与探索, 2022, 16(3): 598-607.
[10]	李想, 杨兴耀, 于炯, 钱育蓉, 郑捷. 基于知识图谱卷积网络的双端推荐算法[J]. 计算机科学与探索, 2022, 16(1): 176-184.
[11]	武家伟, 孙艳春. 融合知识图谱和深度学习方法的问诊推荐系统[J]. 计算机科学与探索, 2021, 15(8): 1432-1440.
[12]	高仰, 刘渊. 融合知识图谱和短期偏好的推荐算法[J]. 计算机科学与探索, 2021, 15(6): 1133-1144.
[13]	舒世泰，李松，郝晓红，张丽平. 知识图谱嵌入技术研究进展[J]. 计算机科学与探索, 2021, 15(11): 2048-2062.
[14]	邢长征，郭亚兰，张全贵，赵宏宝. 融合短文本层级注意力和时间信息的推荐方法[J]. 计算机科学与探索, 2021, 15(11): 2222-2232.
[15]	陈子睿, 王鑫, 王林, 徐大为, 贾勇哲. 开放领域知识图谱问答研究综述[J]. 计算机科学与探索, 2021, 15(10): 1843-1869.