《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (6): 1808-1813.DOI: 10.11772/j.issn.1001-9081.2021091800

• 第十八届CCF中国信息系统及应用大会 • 上一篇    

基于元路径注意力机制的MOOC视频推荐方法

周嘉凡1, 杜岳峰1, 宋宝燕1, 李晓光1(), 赵阿珠2, 肖绪界3   

  1. 1.辽宁大学 信息学院,沈阳 110036
    2.中国人民解放军32286部队,辽宁 铁岭 112600
    3.中国人民解放军92515部队,沈阳 125099
  • 收稿日期:2021-10-22 修回日期:2022-03-11 接受日期:2022-03-14 发布日期:2022-04-15 出版日期:2022-06-10
  • 通讯作者: 李晓光
  • 作者简介:周嘉凡(1996—),女,山西晋城人,硕士研究生,主要研究方向:推荐系统、数据挖掘
    杜岳峰(1986—),男,辽宁沈阳人,讲师,博士,CCF会员,主要研究方向:推荐系统、数据质量管理
    宋宝燕(1965—),女,辽宁铁岭人,教授,博士,CCF会员,主要研究方向:数据库技术
    赵阿珠(1989—)女,辽宁沈阳人,工程师,硕士,主要研究方向:信息系统
    肖绪界(1986—),男,辽宁本溪人,讲师,博士,主要研究方向:数据治理。
  • 基金资助:
    国家重点研发计划项目(2019YFB1406002);国家自然科学基金项目资助项目(U1811261);辽宁省公共舆情与网络安全大数据系统工程实验室基金资助项目(2016-294);辽宁省教育厅科学研究项目(LJKZ0094);辽宁省自然科学基金资助项目(2020-BS-082);辽宁省教育厅青年育苗项目(LQN202010)

MOOC video recommendation method based on meta-path attention mechanism

Jiafan ZHOU1, Yuefeng DU1, Baoyan SONG1, Xiaoguang LI1(), Azhu ZHAO2, Xujie XIAO3   

  1. 1.College of Information,Liaoning University,Shenyang Liaoning 110036,China
    2.PLA 32286 Troop,Tieling Liaoning 112600,China
    3.PLA 92515 Troop,Shenyang Liaoning 125099,China
  • Received:2021-10-22 Revised:2022-03-11 Accepted:2022-03-14 Online:2022-04-15 Published:2022-06-10
  • Contact: Xiaoguang LI
  • About author:ZHOU Jiafan,born in 1996,M. S. candidate. Her research interests include recommendation system,data mining
    DU Yuefeng,born in 1986,Ph. D.,lecturer. His research interests include recommendation system,data quality management
    SONG Baoyan,,born in 1965,Ph. D.,professor. Her research interests include database technologies
    ZHAO Azhu,born in 1989,M. S.,engineer. Her research interests include information system
    XIAO Xujie,born in 1986,Ph. D.,lecturer. His research interests include data governance
  • Supported by:
    National Key Research and Development Program of China(2019YFB1406002);National Natural Science Foundation of China(U1811261);Project of Public Opinion and Network Security Big Data System Engineering Laboratory of Liaoning Province(2016-294);Scientific Research Project of Education Department of Liaoning Province(LJKZ0094);Natural Science Foundation Project of Liaoning Province(2020-BS-082);Youth Nursery Project of Education Department of Liaoning Province(LQN202010)

摘要:

MOOC平台上,一个课程可能存在多个版本的视频,为向学生推荐一个满足学习兴趣的MOOC视频就需要分析学生兴趣与视频内容的关系,为此,提出一种基于元路径注意力机制的视频推荐方法Mrec。一方面,利用异构信息网(HIN)描述学习者和MOOC资源之间的关系,进而使用元路径表达学生和视频之间的交互关系;另一方面,利用注意力机制捕捉学生、视频、元路径的特征对学习兴趣的影响情况。具体来说,Mrec方法包括两层注意力机制:第一层是节点注意力层,通过邻居的特征加权联合节点自身的特征,利用多头注意力得到实体在不同元路径下的特征表示;第二层是路径注意力层,通过融合在不同元路径的指导下学习到的实体的特征表示来捕捉实体在不同兴趣下的特征表示,并将学习到的用户与视频实体输入到多层感知机(MLP)中得到预测分数来进行top-K推荐。在MOOCCube和MOOCdata数据集上进行实验的结果表明,Mrec的点击率、归一化折损累积收益(NDCG)、平均倒数排名(MRR)与受试者工作特征曲线下面积(AUC)均优于对比方法。

关键词: 推荐系统, 异构信息网络, 元路径, 注意力机制, 图神经网络

Abstract:

On the MOOC platform, there may be multiple versions of videos for one course,in order to recommend a MOOC video that satisfies the learning interest of the student,it is necessary to analyze the relationship between student interests and video contents. For this purpose, a video recommendation model named Mrec was proposed based on meta-path attention mechanism. For one thing, the Heterogeneous Information Network (HIN) was used to describe the relationships between learners and MOOC resources, and then meta-path was used to express the interaction between students and videos. For another, the attention mechanism was used to capture the influences of the characteristics of students, videos and meta-paths on learning interest. Specifically, the Mrec model was composed of two layers of attention mechanism: the first layer was the node attention layer, the node own characteristics were weightely combined with neighbor characteristics, and the feature representations of entities under different meta-paths were obtained by multi-head attention; the second layer was the path attention layer, in which the feature representations of entities learned under the guidance of different meta-paths were integrated to capture the feature representations of entities under different interests, and the learned users and video entities were put into Multi-Layer Perceptron (MLP) to obtain the prediction scores for top-K recommendation. Experimental results on MOOCCube and MOOCdata datasets show that Mrec outperforms the comparison methods in terms of Hit Ratio (HR), Normalized Discounted Cumulative Gain (NDCG), Mean Reciprocal Ranking (MRR) and Area Under receiver operating characteristic Curve (AUC).

Key words: recommendation system, Heterogeneous Information Network (HIN), meta-path, attention mechanism, Graph Neural Network (GNN)

中图分类号: