Leveraging semantic features for recommendation: Sentence-level emotion analysis

作者：

Highlights：

• A review-based recommendation model is proposed to separate user reviews into different sentiment orientations.

• A voting mechanism is proposed to directly generate the recommendation set through assigning voting rights according to similarities.

• The effectiveness of separating user reviews into positive and negative feedbacks is validated in the experiments.

摘要

•A review-based recommendation model is proposed to separate user reviews into different sentiment orientations.•A voting mechanism is proposed to directly generate the recommendation set through assigning voting rights according to similarities.•The effectiveness of separating user reviews into positive and negative feedbacks is validated in the experiments.

论文关键词：Personalized recommendation,Text mining,Topic modeling,Sentiment analysis,Cold start

论文评审过程：Received 9 October 2020, Revised 28 December 2020, Accepted 5 February 2021, Available online 11 February 2021, Version of Record 11 February 2021.

论文官网地址：https://doi.org/10.1016/j.ipm.2021.102543