An adaptable fine-grained sentiment analysis for summarization of multiple short online reviews

作者：

Highlights：

• This study presents a novel method in generating online review summaries using an adaptable fine-grained sentiment analysis for short texts.

• A multi-level classification approach is used to build the sentiment classification model.

• To cope up with the weaknesses of both topic models, an extension of LDA and BTM called eBTM is introduced as a technique for extracting aspects.

• Through the sentiment classification and aspect extraction techniques, this study shows the construction of an effective online review summarization technique.

摘要

•This study presents a novel method in generating online review summaries using an adaptable fine-grained sentiment analysis for short texts.•A multi-level classification approach is used to build the sentiment classification model.•To cope up with the weaknesses of both topic models, an extension of LDA and BTM called eBTM is introduced as a technique for extracting aspects.•Through the sentiment classification and aspect extraction techniques, this study shows the construction of an effective online review summarization technique.

论文关键词：Review summarization,Aspect extraction,Sentiment analysis,Short texts,Online reviews

论文评审过程：Received 17 August 2016, Revised 2 March 2017, Accepted 17 March 2017, Available online 3 May 2017, Version of Record 20 July 2017.

论文官网地址：https://doi.org/10.1016/j.datak.2017.03.009