基于知识图谱邻域扩展特征的加权推荐方法
KGTR:WEIGHTED RECOMMENDATION METHOD BASED ON KNOWLEDGE GRAPH NEIGHBORHOOD EXPANSION FEATURE
张晓明 1吴婷婷 1王会勇1
作者信息
- 1. 河北科技大学信息科学与工程学院 河北 石家庄 050000
- 折叠
摘要
知识图谱应用到推荐算法中可以有效降低推荐系统中存在的数据稀疏性和冷启动障碍,因此提出一种基于邻域扩充特征的加权推荐方法(KGTR).在知识图谱卷积网络(KGCN)中加入关系对实体的关联强度,来完成项目特征的邻域表示;将其与扩充用户兴趣特征的模型(RippleNet)结果采用加权的方法结合,并引入KL散度(Kullback-Leibler)作为偏差.在MovieLens-1M和Book-Crossing两个数据集上的实验结果表明,该方法比基线模型在评估指标AUC、F1 和ACC上有提升.
Abstract
The application of the knowledge graph to the recommendation algorithm can effectively reduce the problem of data sparsity and user cold start barriers in the recommendation system.This paper proposes a weighted recommendation method based on the neighborhood expansion feature.This method added the correlation strength between the relationship and the entity to complete the neighborhood representation of the item's features in the knowledge graph convolutional network(KGCN).We combined it with the results of the model(RippleNet)that expanded user's features using a weighted method.The KL(Kullback-Leibler)divergence was introduced as the deviation.The experiment was carried out on two datasets of MovieLens-1M and Book-Crossing.The results show that the proposed method is improved on AUC,F1 and ACC compared with the baseline model.
关键词
知识图谱/推荐算法/邻域特征/三元组表示/关联强度Key words
Knowledge graph/Recommendation algorithm/Neighborhood feature/Triple representation/Correlation strength引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据