Pair-wise ranking based preference learning for points-of-interest recommendation

作者：

Highlights：

• POIs recommendation is addressed from the preference learning perspective.

• A novel pair-wise ranking model is proposed for preference learning.

• Deep neural network is employed for implementing pair-wise ranking.

• A new optimization criterion is proposed for ranking model optimizing.

• A method for building semantic representation of POIs category is proposed.

摘要

•POIs recommendation is addressed from the preference learning perspective.•A novel pair-wise ranking model is proposed for preference learning.•Deep neural network is employed for implementing pair-wise ranking.•A new optimization criterion is proposed for ranking model optimizing.•A method for building semantic representation of POIs category is proposed.

论文关键词：Negative sampling,Neural network,POI recommendation,Pair-wise learning,Semantic representation

论文评审过程：Received 12 May 2020, Revised 15 April 2021, Accepted 20 April 2021, Available online 29 April 2021, Version of Record 8 May 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107069