Two-stage sentiment classification based on user-product interactive information
作者:
Highlights:
• Two-stage classification simulates process of human doing multiple choice questions.
• User-product interactive information judges the sentiment tendency of reviews.
• Both performance and interpretability of the model are improved.
• Model is more effective for more fine-grained sentiment classification task.
摘要
•Two-stage classification simulates process of human doing multiple choice questions.•User-product interactive information judges the sentiment tendency of reviews.•Both performance and interpretability of the model are improved.•Model is more effective for more fine-grained sentiment classification task.
论文关键词:Review document,Sentiment classification,Coarse-to-fine classification,User-product interactive information,Deep learning,Interpretability
论文评审过程:Received 10 June 2019, Revised 28 May 2020, Accepted 30 May 2020, Available online 2 June 2020, Version of Record 4 June 2020.
论文官网地址:https://doi.org/10.1016/j.knosys.2020.106091