Opinion mining for multiple types of emotion-embedded products/services through evolutionary strategy

作者：

Highlights：

• Focus on those reviews on highly emotion-embedded products/services.

• Solve the multiple polarities problem of the same review word for multiple products.

• Propose an evolutionary algorithm for classifying opinions at sentence level.

• Suggest a new SVM based on seven measures for classifying opinion at document level.

摘要

•Focus on those reviews on highly emotion-embedded products/services.•Solve the multiple polarities problem of the same review word for multiple products.•Propose an evolutionary algorithm for classifying opinions at sentence level.•Suggest a new SVM based on seven measures for classifying opinion at document level.

论文关键词：Chinese corpus,Evolutionary strategy,Multiple polarities,Opinion mining,Optimization,Sentiment analysis

论文评审过程：Received 23 July 2017, Revised 14 January 2018, Accepted 15 January 2018, Available online 3 February 2018, Version of Record 3 February 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.01.022