Value Customer Reviews in E-Commerce Platforms using Sentiment Analysis Tactics: A Review

Abstract

With the e-commerce trend, people are inspired to purchase their products online because of their hectic lifestyles. People have to read all the comments to get a clear idea about a product before they purchase it. Reading all the comments to get an idea about a product is impossible, requires a huge effort, and is a waste of time. Sentiment Analysis (SA) provides a way to gain meaningful insights from the huge bulk of customer reviews. It is employed to extract people’s attitudes, emotions, and perceptions towards a product described in the text. In this review paper, a comparative study was conducted by filtering the most relevant literature on SA techniques that have been used in the e-commerce sector. In this study, we have identified issues, gaps, limitations, challenges, and additional requirements in existing SA methods that have been explored by prior researchers. We noted that the lexicon-based method used pre-compiled sentiments and that the machine learning-based method was more intelligent than the lexicon approaches. The hybrid approach was a combination of lexicon and machine learning approaches. The requirement for high computational power prevents the wide use of hybrid approaches. However, reliability and performance efficiency are higher with the hybrid approach. We note that previous studies have not proposed an efficient system for using SA in the e-commerce sector. Therefore, we recommend a novel approach that combines both lexicon and machine learning techniques.