《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (8): 2358-2363.DOI: 10.11772/j.issn.1001-9081.2022091336

• 第十九届CCF中国信息系统及应用大会 • 上一篇    

面向个性化课程推荐的分层分期注意力网络模型

刘源1,2, 董永权1,2(), 贾瑞1, 杨昊霖1   

  1. 1.江苏师范大学 计算机科学与技术学院,江苏 徐州 221116
    2.徐州市云计算工程技术研究中心(江苏师范大学),江苏 徐州 221116
  • 收稿日期:2022-09-06 修回日期:2022-09-21 接受日期:2022-09-26 发布日期:2022-09-29 出版日期:2023-08-10
  • 通讯作者: 董永权
  • 作者简介:刘源(1997—),男, 山东淄博人,硕士研究生,CCF会员,主要研究方向:推荐系统、深度学习
    贾瑞(1999—),女,山西临汾人,硕士研究生,主要研究方向:知识追踪、深度学习
    杨昊霖(1997—),男,河南商丘人,硕士研究生,主要研究方向:数据集成、多真值发现。
  • 基金资助:
    国家自然科学基金资助项目(61872168);江苏省教育科学“十四五”规划课题(D/2021/01/112);江苏师范大学科研与实践创新项目(2022XKT1544)

Hierarchical and phased attention network model for personalized course recommendation

Yuan LIU1,2, Yongquan DONG1,2(), Rui JIA1, Haolin YANG1   

  1. 1.School of Computer Science and Technology,Jiangsu Normal University,Xuzhou Jiangsu 221116,China
    2.Xuzhou Cloud Computing Engineering Technology Research Center (Jiangsu Normal University),Xuzhou Jiangsu 221116,China
  • Received:2022-09-06 Revised:2022-09-21 Accepted:2022-09-26 Online:2022-09-29 Published:2023-08-10
  • Contact: Yongquan DONG
  • About author:LIU Yuan, born in 1997, M. S. candidate. His research interests include recommender system, deep learning.
    JIA Rui, born in 1999, M. S. candidate. Her research interests include knowledge tracking, deep learning.
    YANG Haolin, born in 1997, M. S. candidate. His research interests include data integration, multi-truth value discovery.
  • Supported by:
    National Natural Science Foundation of China(61872168);Jiangsu Education Science "the 14th Five Year Plan" Program(D/2021/01/112);Scientific Research and Practical Innovation Project of Jiangsu Normal University(2022XKT1544)

摘要:

随着大规模在线开放课程(MOOC)平台的广泛使用,需要一种有效的方法为用户推荐个性化课程。针对现有的课程推荐方法通常利用课程学习记录为用户的学习兴趣建立整体的静态表示,但忽略了学习兴趣动态变化与用户短期学习兴趣的问题,提出一种分层分期的注意力网络(HPAN)进行个性化课程推荐。该网络的第1层利用注意力网络得到用户的长短期学习兴趣,第2层将用户的长短期学习兴趣和短期交互序列相结合并通过注意力网络得到用户的兴趣向量;然后计算用户兴趣向量与每个课程向量的偏好值,据此为用户进行课程推荐。在XuetangX(学堂在线)公开数据集上的实验结果表明,与次优的序列分层注意力网络(SHAN)模型相比,HPAN模型的Recall@5提高了12.7%,与FPMC(Factorizing Personalized Markov Chains)模型相比,它的MRR@20提高了15.6%。HPAN模型的推荐效果优于对比模型,可解决实际的个性化课程推荐。

关键词: 推荐系统, 课程推荐, 注意力机制, 个性化推荐, 长短期兴趣

Abstract:

With the widespread applications of Massive Open Online Courses (MOOCs) platforms, an effective method is needed for personalized course recommendation. In view of the existing course recommendation methods, which usually use the course learning records to establish the overall static representation for users’ learning interests, while ignoring the dynamic changes of learning interests and users’ short-term learning interests, a Hierarchical and Phased Attention Network (HPAN) was proposed to carry out personalized course recommendation. In the first layer of the network, the attention network was used to obtain the user’s long- and short-term learning interests. In the second layer of the network, the user’s long- and short-term learning interests and short-term interaction sequence were combined to obtain the user’s interest vector through the attention network, then the preference value of the user’s interest vector with each course vector was calculated, and courses were recommended for the user according to the result. Experimental results on public dataset XuetangX show that, compared with the second best SHAN (Sequential Hierarchical Attention Network) model, HPAN model has the Recall@5 increased by 12.7%; compared with FPMC (Factorizing Personalized Markov Chains) model, HPAN model has the MRR@20 increased by 15.6%. HPAN model has better recommendation effect than the comparison models, and can be used for practical personalized course recommendation.

Key words: recommender system, course recommendation, attention mechanism, personalized recommendation, long- and short-term interest

中图分类号: