Multi-objective item evaluation for diverse as well as novel item recommendations

作者：

Highlights：

• All item pair based Bhat_sim similarity model enhances the rating prediction accuracy.

• Genetic item order and frequency aware multiparent crossover foster search capability.

• Trade-off between precision and topic diversity filters diverse as well as novel items.

• NewCrossPMOEA increases mean precision, diversity, and novelty of recommendation list.

摘要

•All item pair based Bhat_sim similarity model enhances the rating prediction accuracy.•Genetic item order and frequency aware multiparent crossover foster search capability.•Trade-off between precision and topic diversity filters diverse as well as novel items.•NewCrossPMOEA increases mean precision, diversity, and novelty of recommendation list.

论文关键词：Recommendation system,Collaborative filtering,Multi-objective optimization,Non-linear similarity,Bhattacharyya coefficient

论文评审过程：Received 1 May 2019, Revised 30 July 2019, Accepted 31 July 2019, Available online 1 August 2019, Version of Record 9 August 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.112857