A comprehensive survey on sentiment analysis: Approaches, challenges and trends

作者：

Highlights：

• Sentiment analysis is constantly evolving through approaches, data and models.

• The paper provides an unprecedented and comprehensive survey on sentiment analysis.

• Traditional and recent models are discussed, compared and classified.

• Pointing out the reasons to select the proper model for sentiment analysis.

• The paper summarizes the sentiment analysis models to monitor future trends.

摘要

•Sentiment analysis is constantly evolving through approaches, data and models.•The paper provides an unprecedented and comprehensive survey on sentiment analysis.•Traditional and recent models are discussed, compared and classified.•Pointing out the reasons to select the proper model for sentiment analysis.•The paper summarizes the sentiment analysis models to monitor future trends.

论文关键词：Sentiment analysis,Opinion mining,Machine learning,Lexicon-based,Sentiment classification,Deep learning

论文评审过程：Received 1 July 2020, Revised 25 March 2021, Accepted 10 May 2021, Available online 14 May 2021, Version of Record 18 May 2021.

论文官网地址：https://doi.org/10.1016/j.knosys.2021.107134