《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (8): 2485-2493.DOI: 10.11772/j.issn.1001-9081.2025070864

• 数据科学与技术 • 上一篇    下一篇

融合多行为建模与强化学习的学习路径推荐模型

陈鹏宇1, 田保军1,2(), 赵利畅1, 房建东2,3   

  1. 1.内蒙古工业大学 智能科学与技术学院,呼和浩特 010080
    2.内蒙古自治区智能感知与系统工程重点实验室(内蒙古工业大学),呼和浩特 010080
    3.内蒙古工业大学 信息工程学院,呼和浩特 010080
  • 收稿日期:2025-07-31 修回日期:2025-10-11 接受日期:2025-10-14 发布日期:2025-11-05 出版日期:2026-08-10
  • 通讯作者: 田保军
  • 作者简介:陈鹏宇(2000—),男,重庆人,硕士研究生,主要研究方向:推荐系统、机器学习
    田保军(1971—),男,内蒙古呼和浩特人,教授,硕士,主要研究方向:机器学习、推荐系统
    赵利畅(2001—),男,山东德州人,硕士研究生,主要研究方向:机器学习、推荐系统
    房建东(1966—),女,内蒙古呼和浩特人,教授,博士,主要研究方向:信息处理、智能控制。
  • 基金资助:
    内蒙古自治区科技计划项目(2025KYPT0088);内蒙古自治区自然科学基金资助项目(2023LHMS06008);鄂尔多斯市2025年科技突围工程“揭榜挂帅”重大项目“人工智能+”专项(JB20251436)

Learning path recommendation model integrating multi-behavior modeling and reinforcement learning

Pengyu CHEN1, Baojun TIAN1,2(), Lichang ZHAO1, Jiandong FANG2,3   

  1. 1.College of Intelligent Science and Technology,Inner Mongolia University of Technology,Hohhot Inner Mongolia 010080,China
    2.Inner Mongolia Key Laboratory of Intelligent Perception and System Engineering (Inner Mongolia University of Technology),Hohhot Inner Mongolia 010080,China
    3.College of Information Engineering,Inner Mongolia University of Technology,Hohhot Inner Mongolia 010080,China
  • Received:2025-07-31 Revised:2025-10-11 Accepted:2025-10-14 Online:2025-11-05 Published:2026-08-10
  • Contact: Baojun TIAN
  • About author:CHEN Pengyu, born in 2000, M. S. candidate. His research interests include recommendation system, machine learning.
    ZHAO Lichang, born in 2001, M. S. candidate. His research interests include machine learning, recommendation system.
    FANG Jiandong, born in 1966, Ph. D., professor. Her research interests include information processing, intelligent control.
  • Supported by:
    Inner Mongolia Scientific and Technological Program(2025KYPT0088);Natural Science Foundation of Inner Mongolia Autonomous Region(2023LHMS06008);“Artificial Intelligence+” Special Project under “Unveiling and Commanding” Major Initiatives of Ordos City’s 2025 Technology Breakthrough Engineering(JB20251436)

摘要:

针对学习路径推荐任务中的交互数据稀疏和学习资源规划不合理问题,提出一种融合多行为建模与强化学习(RL)的学习路径推荐模型LPRMMMRL。首先,将用户行为数据分类并作为实体节点融入课程知识图谱,使用带有注意力机制的图卷积网络(GCN)捕捉多源异构特征;其次,从行为与学习资源维度设计3种数据增强策略,并采用对比学习方法对增强后的数据进行表征学习且把它们作为补充信息纳入RL;最后,充分考虑用户的个体差异,通过深度Q网络(DQN)构建个性化学习路径推荐模型,并设计知识点与路径双重奖励函数机制以引导模型收敛。在真实数据集MOOPer和MOOCCubeX的4门课程上,将LPRMMMRL与LPG(Learning Path Generator)、KTKDM(Knowledge Tracing based Knowledge Demand Model)等学习路径推荐模型进行对比。实验结果表明,LPRMMMRL的精确率、召回率和F1分数在MOOPer数据集上分别至少提升了5.53%和5.74%,在MOOCCubeX数据集上分别至少提升了4.56%和3.41%。此外,通过参数敏感性分析实验与消融实验结果,验证了所提模型的有效性。

关键词: 学习路径推荐, 多行为建模, 强化学习, 异构知识图谱, 图卷积网络

Abstract:

To address issues of sparse interaction data and unreasonable learning resource planning in learning path recommendation, a Learning Path Recommendation Model integrating Multi-behavior Modeling and Reinforcement Learning (RL), LPRMMMRL, was proposed. First, user behavior data were classified and fused into a course knowledge graph as entity nodes, and a Graph Convolutional Network (GCN) with attention mechanism was used to capture multi-source heterogeneous features. Second, three data augmentation strategies were designed from both behavioral and learning resource perspectives, and contrastive learning was applied to carry out representation learning to the augmented data, which were integrated into RL as supplementary information. Finally, after fully considering individual differences of the users, a personalized learning path recommendation model was built using a Deep Q-Network (DQN), and a dual reward function of both knowledge points and paths was designed to guide model convergence. Experiments were conducted for comparing LPRMMMRL with learning path recommendation models such as LPG (Learning Path Generator), KTKDM (Knowledge Tracing based Knowledge Demand Model) on four real-world courses from MOOPer and MOOCCubeX datasets. The results show that LPRMMMRL achieves at least 5.53% and 5.74% improvements in precision, recall, and F1-score on the MOOPer dataset, respectively, and at least 4.56% and 3.41% improvements on the MOOCCubeX dataset, respectively. Besides, parameter sensitivity analysis and ablation studies validate the effectiveness of the proposed model.

Key words: learning path recommendation, multi-behavior modeling, Reinforcement Learning (RL), heterogeneous knowledge graph, Graph Convolutional Network (GCN)

中图分类号: