Computer Engineering and Applications ›› 2022, Vol. 58 ›› Issue (5): 94-103.DOI: 10.3778/j.issn.1002-8331.2109-0126

• Big Data and Cloud Computing • Previous Articles     Next Articles

Recommendation Algorithm Combining Knowledge Graph and Attention Mechanism

TANG Hong, FAN Sen, TANG Fan , ZHU Longjiao   

  1. 1.School of Communication and Information Engineering, Chongqing University of Posts and Communications, Chongqing 400065, China
    2.Chongqing Key Lab of Mobile Communications Technology, Chongqing University of Posts and Communications, Chongqing 400065, China
  • Online:2022-03-01 Published:2022-03-01



  1. 1.重庆邮电大学 通信与信息工程学院,重庆 400065
    2.重庆邮电大学 移动通信技术重庆市重点实验室,重庆 400065

Abstract: In order to solve the problem of information overload, this paper proposes a recommendation model that combines knowledge graph and attention mechanism. First of all, in the model, embedding the knowledge graph as auxiliary information can alleviate the data sparseness and cold start problems of the traditional recommendation algorithm, and bring interpretability to the final recommendation result. Secondly, in order to improve the recommendation accuracy and capture the dynamic changes of user interests, combined with the neural network in deep learning and the attention mechanism to generate user-adaptive representations, plus dynamic factors to better capture user dynamic changes in interest, using multiple layers perceptron makes scoring predictions for items. Finally, the simulation verification is performed on the MovieLens-latest-small movie dataset and the Douban dataset. The results show that the model for TOP-K list recommendation has better recommendation performance than other algorithms.

Key words: knowledge graph, neural network, attention mechanism, multilayer perceptron, recommendation algorithm

摘要: 为了解决信息过载问题,提出了一种融合知识图谱与注意力机制的推荐模型。在该模型中,将知识图谱作为辅助信息进行嵌入,可以缓解传统推荐算法数据稀疏和冷启动问题,并且给推荐结果带来可解释性。为了提升推荐准确率以及捕捉用户兴趣的动态变化,再结合深度学习中的神经网络以及注意力机制生成用户自适应表示,加上动态因子来更好地捕捉用户动态兴趣变化,使用多层感知机对项目进行评分预测。在MovieLens-latest-small电影数据集和豆瓣数据集进行仿真验证,结果表明该模型进行TOP-K列表电影推荐相比于其他算法拥有更好的推荐性能。

关键词: 知识图谱, 神经网络, 注意力机制, 多层感知机, 推荐算法