ClustPTF: Clustering-based parallel tensor factorization for the diverse multi-criteria recommendation

作者：

Highlights：

• The ClustPTF remarkably improves recommendation diversity and predictive performance.

• Sentiment analysis is used to reduce model sparsity by complementing missing ratings.

• Two data structures of multi-criteria preferences are introduced for clustering them.

• Experimental results show ClustPTF’s potential on a near real-time recommendation.

摘要

•The ClustPTF remarkably improves recommendation diversity and predictive performance.•Sentiment analysis is used to reduce model sparsity by complementing missing ratings.•Two data structures of multi-criteria preferences are introduced for clustering them.•Experimental results show ClustPTF’s potential on a near real-time recommendation.

论文关键词：Clustering,Recommendation diversity,Multi-criteria Recommender system,Sentiment analysis,Parallel Tensor factorization

论文评审过程：Received 17 October 2020, Revised 18 January 2021, Accepted 11 March 2021, Available online 24 March 2021, Version of Record 13 April 2021.

论文官网地址：https://doi.org/10.1016/j.elerap.2021.101041