KLGCN: Knowledge graph-aware Light Graph Convolutional Network for recommender systems

作者：

Highlights：

• Explorations of feature transformation and nonlinear activation in KG-aware recommendation.

• Parallel aggregating features on two source graphs in a light and effective way.

• Importance of users’ personalized interests and neighbor roles.

• Satisfactory performance results on three widely used, real-world datasets.

摘要

•Explorations of feature transformation and nonlinear activation in KG-aware recommendation.•Parallel aggregating features on two source graphs in a light and effective way.•Importance of users’ personalized interests and neighbor roles.•Satisfactory performance results on three widely used, real-world datasets.

论文关键词：Recommender systems,Knowledge graph,Graph convolutional network,Collaborative filtering

论文评审过程：Received 4 July 2021, Revised 3 January 2022, Accepted 4 January 2022, Available online 29 January 2022, Version of Record 10 February 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.116513