《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (4): 1057-1064.DOI: 10.11772/j.issn.1001-9081.2021071255

• CCF第36届中国计算机应用大会 (CCF NCCA 2021) • 上一篇    

基于无采样协作知识图网络的推荐系统

蒋雯静1,2, 熊熙1,2,3(), 李中志1,2, 李斌勇1,2   

  1. 1.成都信息工程大学 网络空间安全学院,成都 610255
    2.先进密码技术与系统安全四川省重点实验室(成都信息工程大学),成都 610225
    3.四川大学 空天科学与工程学院,成都 610065
  • 收稿日期:2021-07-16 修回日期:2021-08-20 接受日期:2021-08-25 发布日期:2021-08-20 出版日期:2022-04-10
  • 通讯作者: 熊熙
  • 作者简介:蒋雯静(1995—),女,四川乐山人,硕士研究生,主要研究方向:推荐系统、知识图谱
    李中志(1973—),男,四川泸州人,副教授,硕士,CCF会员,主要研究方向:数据挖掘
    李斌勇(1982—),男,四川绵阳人,讲师,博士,CCF会员,主要研究方向:工业智能化、数据挖掘。
  • 基金资助:
    国家自然科学基金资助项目(81901389);教育部“春晖计划”合作科研项目(2019);中国博士后科学基金资助项目(2019M653400)

Recommendation system based on non-sampling collaborative knowledge graph network

Wenjing JIANG1,2, Xi XIONG1,2,3(), Zhongzhi LI1,2, Binyong LI1,2   

  1. 1.School of Cybersecurity,Chengdu University of Information Technology,Chengdu Sichuan 610225,China
    2.Advanced Cryptography and System Security Key Laboratory of Sichuan Province(Chengdu University of Information Technology),Chengdu Sichuan 610225,China
    3.School of Aeronautics and Astronautics,Sichuan University,Chengdu Sichuan 610065,China
  • Received:2021-07-16 Revised:2021-08-20 Accepted:2021-08-25 Online:2021-08-20 Published:2022-04-10
  • Contact: Xi XIONG
  • About author:JIANG Wenjing, born in 1995, M. S. candidate. Her research interests include recommendation system, knowledge graph.
    LI Zhongzhi, born in 1973, M. S., associate professor. His research interests include data mining.
    LI Binyong, born in 1982, Ph. D., lecturer. His research interests include industrial intelligence, data mining.
  • Supported by:
    National Natural Science Foundation of China(81901389);“Chunhui Program” Cooperative Scientific Research Project of Ministry of Education (2019), China Postdoctoral Science Foundation(2019M653400)

摘要:

知识图谱(KG)可以通过高效组织海量数据实现信息的有效抽取,因而基于知识图谱的推荐方法得到了广泛的研究和应用。针对图神经网络在知识图谱建模中的采样误差问题,提出了一种无采样协作知识图网络(NCKN)的方法。首先,设计了无采样知识传播模块,通过在单个卷积层使用不同大小的线性聚合器来捕捉深层次的信息,实现高效的无采样预计算;然后,为了区分邻居节点贡献度,在传播过程中引入注意力机制;最后,协作传播模块将知识嵌入同用户交互中的协作信号相结合,以更好地描述用户偏好。基于三个真实数据集,评估了NCKN在CTR预测和Top-k预测中的性能。实验结果表明,与主流算法RippleNet、知识图卷积神经网络(KGCN)相比,NCKN在CTR预测中的准确率平均分别提升了2.71%、4.60%;Top-k预测中,NCKN的准确率平均分别提升了5.26%、3.91%。所提方法不仅解决了图神经网络在知识图谱建模中的采样误差问题,且提升了推荐模型的准确率。

关键词: 知识图谱, 推荐系统, 采样策略, 图神经网络, 注意力机制

Abstract:

Knowledge Graph (KG) can effectively extract information by efficiently organizing massive data. Therefore, recommendation methods based on knowledge graph have been widely studied and applied. Aiming at the sampling error problem of graph neural network in knowledge graph modeling, a method of Non-sampling Collaborative Knowledge graph Network (NCKN) was proposed. Firstly, a non-sampling knowledge dissemination module was designed, in which linear aggregators with different sizes were used in a single convolutional layer to capture deep-level information and achieve efficient non-sampling pre-computation. Then, in order to distinguish the contribution degrees of neighbor nodes, attention mechanism was introduced in the dissemination process. Finally, the collaboration signal of user interaction and knowledge embedding were combined in the collaborative dissemination module to better describe user preferences. Based on three real datasets, the performance of NCKN in CTR (Click Through Rate) prediction and Top-k was evaluated. The experimental results show that compared with the mainstream algorithms RippleNet (Ripple Network) and KGCN (Knowledge Graph Convolutional Network), the accuracy of NCKN in CTR prediction increases by 2.71% and 4.60%, respectively; in the Top-k forecast, prediction, the accuracy of NCKN increases by 5.26% and 3.91% on average respectively. The proposed method not only solves the sampling error problem of graph neural network in knowledge map modeling, but also improves the accuracy of the recommended model.

Key words: Knowledge Graph (KG), recommendation system, sampling strategy, graph neural network, attention mechanism

中图分类号: