Automatic construction of domain-specific sentiment lexicon for unsupervised domain adaptation and sentiment classification
作者：
Highlights：
• An unsupervised approach for sentiment analysis in unknown domains is proposed.
• The approach generates a domain-specific sentiment lexicon.
• Sentiment information from a domain is adapted using a Domain Independent Lexicon.
• The weights of the first layer of an MLP are set using the resulting polarities.
• The experiments on multi-domain datasets show the effectiveness of the approach.
摘要
•An unsupervised approach for sentiment analysis in unknown domains is proposed.•The approach generates a domain-specific sentiment lexicon.•Sentiment information from a domain is adapted using a Domain Independent Lexicon.•The weights of the first layer of an MLP are set using the resulting polarities.•The experiments on multi-domain datasets show the effectiveness of the approach.
论文关键词：Sentiment analysis,Domain adaptation,Domain independent lexicon,Multilayer perceptron
论文评审过程：Received 10 January 2020, Revised 15 August 2020, Accepted 1 September 2020, Available online 7 November 2020, Version of Record 21 January 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2020.106423
