A social recommendation method based on an adaptive neighbor selection mechanism
作者：
Highlights：
• An adaptive neighbor selection mechanism is proposed for recommender systems.
• The proposed method incorporates social information into recommendation process.
• The quality of predicted ratings is evaluated using a reliability measure.
• A new confidence model is proposed to identify useless users from neighbors set.
• Experiments show that our method outperforms several state-of-the-art methods.
摘要
•An adaptive neighbor selection mechanism is proposed for recommender systems.•The proposed method incorporates social information into recommendation process.•The quality of predicted ratings is evaluated using a reliability measure.•A new confidence model is proposed to identify useless users from neighbors set.•Experiments show that our method outperforms several state-of-the-art methods.
论文关键词：Recommender systems,Adaptive neighbor selection,Confidence,Reliability,Trust
论文评审过程：Received 9 November 2016, Revised 30 January 2017, Accepted 10 March 2017, Available online 18 March 2017, Version of Record 14 May 2018.
论文官网地址：https://doi.org/10.1016/j.ipm.2017.03.002
