动态知识图谱与强化学习下职教资源个性化推荐算法

Personalized Recommendation Algorithm for Vocational Education Resources Based on Dynamic Knowledge Graph and Reinforcement Learning

摘要

针对在职教资源个性化推荐算法采用的混沌映射构建推荐学习模型易受维度灾难影响, 导致 NDCG(Normalized Discounted Cumulative Gain)归一化折损增益不理想的问题, 对基于动态知识图谱与强化学习的职教资源个性化推荐算法进行优化设计。构建多源异构数据的职教资源动态知识图谱表征, 引入时序感知的嵌入表征向量得到动态嵌入函数; 通过图注意力网络聚合实体的邻域信息, 更新节点表征, 从而高度集成动态知识图谱的结构信息; 采用 Actor-Critic 架构将目标函数作为最大化累计期望回报, 引入熵正则化, 在满足推荐需求平衡的基础上得到复合奖励式; 引入代价敏感的搜索算法与强化学习策略相结合, 生成启发式评估函数, 通过重采样机制在动态知识图谱中寻找有效资源进行路径插入, 规划自适应推荐路径, 完成职教资源个性化推荐。实验结果表明, 研究算法在不同类型资源下的推荐 NDCG 归一化折损增益始终高于 0.8, 在排序时更能考虑物品的相关性和位置因素, 将用户更可能感兴趣的物品排在更靠前的位置, 推荐质量良好, 未受浏览筛选迟滞作用影响出现需求不匹配问题, 具有推荐显著一致性。

Abstract

In the design of personalized recommendation algorithms for vocational education resources, chaotic mapping is commonly used to construct recommendation learning models. However, such models are susceptible to the curse of dimensionality, leading to suboptimal NDCG (Normalized Discounted Cumulative Gain). To address this, an optimized design of a personalized recommendation algorithm is proposed for vocational education resources based on dynamic knowledge graph and reinforcement learning. Specifically, a dynamic knowledge graph representation is constructed from multi-source heterogeneous vocational education resource data, and a temporal-aware embedding representation vector is introduced to derive dynamic embedding functions. A graph attention network is employed to aggregate neighborhood information of entities, updating node representations to deeply integrate structural information of the dynamic knowledge graph. An Actor-Critic framework is adopted to maximize cumulative expected rewards, and entropy regularization is introduced to achieve a composite reward function that balances recommendation objectives. By combining a cost-sensitive search algorithm with the reinforcement learning strategy, a heuristic evaluation function is generated. A resampling mechanism is applied to locate effective resources within the dynamic knowledge graph for path insertion, enabling adaptive recommendation path planning and achieving personalized recommendation of vocational education resources. Experimental results demonstrate that the proposed algorithm consistently achieves NDCG values above 0.8 across different types of resources. It effectively accounts for item relevance and ranking positions, placing items of higher user interest in more prominent positions. The recommendation quality remains high without being affected by browsing and filtering delays or demand mismatches, exhibiting significant recommendation consistency.

关键词

动态知识图谱 / 强化学习 / 职教资源 / 资源推荐 / 推荐算法

Key words

dynamic knowledge graph / reinforcement learning / vocational education resources / resource recommendation / recommendation algorithm

参考文献

[1] 刘莹, 杨淑萍, 张治国. 基于 Spark 平台的并行化谱聚类算法的在线学习资源推荐[J]. 济南大学学报(自然科学版), 2024, 38(4): 456-461.

[2] LIU Y, YANG S P, ZHANG Z G. Online Learning Resource Recommendation by Using Parallel Spectral Clustering Algorithm Based on Spark Platform[J]. Journal of University of Jinan (Science and Technology), 2024, 38(4): 456-461.

[3] 陈斌, 顾珑. 基于知识图谱的设备电子信息资源精准推荐算法[J]. 吉林大学学报(信息科学版), 2025, 43(6): 1388-1396.

[4] CHEN B, GU L. Precise Recommendation Algorithm for Information Resources of Equipment Electronic Based on Knowledge Graph[J]. Journal of Jilin University (Information Science Edition), 2025, 43(6): 1388-1396.

[5] 李蒙蒙. 基于用户画像与关联规则的图书馆资源组合推荐算法[J]. 中国科技资源导刊, 2023, 55(2): 104-110.

[6] LI M M. Library Resource Combination Recommendation Algorithm Based on User Portrait and Association Rules[J]. China Science & Technology Resources Review, 2023, 55(2): 104-110.

[7] 陆涵, 王宇琴, 陈姝彤, 等. 一种基于改进鸽群算法的个性化学习资源推荐方法[J]. 南京晓庄学院学报, 2025, 41(6): 110-117.

[8] LU H, WANG Y Q, CHEN S T, et al. A Personalized Learning Resource Recommendation Method Based on an Improved Pigeon-Inspired Optimization Algorithm[J]. Journal of Nanjing Xiaozhuang University, 2025, 41(6): 110-117.

[9] 仇焕青, 陈曙光, 龚芝, 等. 基于鲸鱼优化算法改进长短期记忆神经网络的资源推荐[J]. 济南大学学报(自然科学版), 2023, 37(3): 309-315.

[10] QIU H Q, CHEN S G, GONG Z, et al. Resource Recommendation Based on Long Short Term Memory Neural Network Improved by Using Whale Optimization Algorithm[J]. Journal of University of Jinan (Science and Technology), 2023, 37(3): 309-315.

[11] 随晓文. 基于标签相似度的数字化书籍资源智能推荐算法[J]. 吉林大学学报(信息科学版), 2024, 42(3): 516-521.

[12] SUI X W. Intelligent Recommendation Algorithm of Digital Book Resources Based on Tag Similarity[J]. Journal of Jilin University (Information Science Edition), 2024, 42(3): 516-521.

[13] 程娟娟, 宋彪, 李微. 基于 Vanilla 算法的数字化资源在线推荐算法[J]. 计算机仿真, 2023, 40(9): 324-327, 338.

[14] CHENG J J, SONG B, LI W. The Online Recommendation Algorithm of Digital Resources Based on Vanilla Algorithm[J]. Computer Simulation, 2023, 40(9): 324-327, 338.

[15] 孙浩, 周力青. 基于 IMD 与 Canopy 的非均匀资源推荐算法[J]. 计算机仿真, 2023, 40(10): 477-481.

[16] SUN H, ZHOU L Q. Non-Uniform Resource Recommendation Algorithm Based on Improved Matrix Factorization and Canopy[J]. Computer Simulation, 2023, 40(10): 477-481.

[17] 扁晓倩. 基于协同过滤算法的英语资源库信息推荐模型[J]. 信息技术, 2023(2): 12-16, 23.

[18] BIAN X Q. Information Recommendation Model of English Resource Base Based on Collaborative Filtering Algorithm[J]. Information Technology, 2023(2): 12-16,23.

[19] 朱海萍, 王子瑜, 赵成成, 等. 基于时空多粒度兴趣建模的学习资源推荐方法[J]. 计算机研究与发展, 2025, 62(8): 1884-1901.

[20] ZHU H P, WANG Z Y, ZHAO C C, et al. Learning Resource Recommendation Method Based on Spatio-Temporal Multi-Granularity Interest Modeling[J]. Journal of Computer Research and Development, 2025, 62(8): 1884-1901.

[21] 甘沛露, 宋一豪, 朱晓雄, 等. 融入概率矩阵分解模型的改进二部图推荐算法[J]. 运筹与管理, 2025, 34(1): 1-7.

[22] GAN P L, SONG Y H, ZHU X X, et al. Improved Bipartite Graph Recommendation Algorithm Integrating Probability Matrix Factorization Model[J]. Operations Research and Management Science, 2025, 34(1): 1-7.

[23] 王禾玲. 基于汤普森采样算法的智慧教学平台推荐研究[J]. 自动化与仪器仪表, 2025(12): 177-180, 185.

[24] WANG H L. Smart Teaching Platform Recommendation Based on Thompson Sampling Algorithm[J]. Automation & Instrumentation, 2025(12): 177-180, 185.

[25] 汤志康, 武毓琦, 李春英, 等. 基于知识图谱卷积网络的学习资源推荐[J]. 计算机工程, 2024, 50(9): 153-160.

[26] TANG Z K, WU Y Q, LI C Y, et al. Recommendation of Learning Resource Based on Knowledge Graph Convolutional Network[J]. Computer Engineering, 2024, 50(9): 153-160.

[27] 易晓宇, 易绵竹. 基于兴趣信息深度融合的网络图书资源推荐[J]. 吉林大学学报(工学版), 2024, 54(12): 3614-3619.

[28] YI X Y, YI M Z. Network Book Resource Recommendation Based on Deep Fusion of Interest Information[J]. Journal of Jilin University (Engineering and Technology Edition), 2024, 54(12): 3614-3619.

[29] 陆鹏, 韦曼莉. 算法推荐技术赋能铸牢中华民族共同体意识宣传教育的逻辑机理与实践进路[J]. 北方民族大学学报(哲学社会科学版), 2025(2): 24-34.

[30] LU P, WEI M L. The Logical Mechanism and Practical Approach of Empowering the Promotion and Education of Forging the Sense of Community for Chinese Nation through Algorithm Recommendation Technology[J]. Journal of North Minzu University (Philosophy and Social Science), 2025(2): 24-34.

[31] 单涛, 钱琪杰, 景慎旗, 等. 基于隐语义模型的物联网资源发现算法[J]. 数据采集与处理, 2023, 38(6): 1369-1379.

[32] SHAN T, QIAN Q J, JING S Q, et al. IoT Resource Discovery Algorithm Based on Latent Factor Model[J]. Journal of Data Acquisition & Processing, 2023, 38(6): 1369-1379.

基金资助

教育部产学合作协同育人基金资助项目(240703817125047)