SentiDraw: Using star ratings of reviews to develop domain specific sentiment lexicon for polarity determination
作者：
Highlights：
• Polarity determination using proposed domain specific lexicon delivers better performance than other lexicon based methods
• Using star rating distribution is a novel lexicon building approach for sentiment analysis
• A robust set of experiments have been performed in this study by comparing this method with similar methods that use review labels for training lexicons across several datasets.
• Hybrid method using ML also illustrates the superiority of this method for building domain specific lexicons
• Experiments on popular datasets like Cornell movie reviews dataset and Large movie reviews dataset have also been performed to benchmark against other studies and demonstrate the superior performance of this method.
摘要
•Polarity determination using proposed domain specific lexicon delivers better performance than other lexicon based methods•Using star rating distribution is a novel lexicon building approach for sentiment analysis•A robust set of experiments have been performed in this study by comparing this method with similar methods that use review labels for training lexicons across several datasets.•Hybrid method using ML also illustrates the superiority of this method for building domain specific lexicons•Experiments on popular datasets like Cornell movie reviews dataset and Large movie reviews dataset have also been performed to benchmark against other studies and demonstrate the superior performance of this method.
论文关键词：Polarity determination,Sentiment analysis,Sentiment lexicon,Classification of reviews,SentiDraw,Star rating
论文评审过程：Received 6 January 2020, Revised 15 August 2020, Accepted 9 October 2020, Available online 28 October 2020, Version of Record 28 October 2020.
论文官网地址：https://doi.org/10.1016/j.ipm.2020.102412
