Learning bilingual sentiment lexicon for online reviews
作者：
Highlights：
• Sentiment analysis is an essential task for analyzing online reviews. A Bilingual sentiment lexicon induction approach (BSLI) is proposed to generate bilingual sentiment lexicons for online reviews.
• Our BSLI can propagate lexical resources of a dominant language to the other low-resource language using word embedding techniques.
• Compared to two general lexicons and three alternative methods, our BSLI generate lexicons with better accuracy and coverage.
摘要
•Sentiment analysis is an essential task for analyzing online reviews. A Bilingual sentiment lexicon induction approach (BSLI) is proposed to generate bilingual sentiment lexicons for online reviews.•Our BSLI can propagate lexical resources of a dominant language to the other low-resource language using word embedding techniques.•Compared to two general lexicons and three alternative methods, our BSLI generate lexicons with better accuracy and coverage.
论文关键词：Online review,Bilingual sentiment lexicon induction,Bilingual word embedding,Sentiment analysis
论文评审过程：Received 5 November 2020, Revised 10 February 2021, Accepted 11 February 2021, Available online 24 February 2021, Version of Record 13 March 2021.
论文官网地址：https://doi.org/10.1016/j.elerap.2021.101037
