Understanding patient reviews with minimum supervision

作者：

Highlights：

• Jase, a new neural model for joint sentiment classification and aspect extraction, does not require aspect-level annotations for training.

• Jase learns document representations which capture both sentiment and aspect information in the embedding space.

• Jase extracts semantically more coherence aspect topics compared to traditional and neural topic models.

摘要

•Jase, a new neural model for joint sentiment classification and aspect extraction, does not require aspect-level annotations for training.•Jase learns document representations which capture both sentiment and aspect information in the embedding space.•Jase extracts semantically more coherence aspect topics compared to traditional and neural topic models.

论文关键词：Sentiment analysis,Aspect extraction,Patient reviews

论文评审过程：Received 29 September 2020, Revised 8 July 2021, Accepted 16 August 2021, Available online 1 September 2021, Version of Record 10 September 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102160