Sentiment Analysis of Product Reviews Using Deep Learning and Transformer Models: A Comparative Study

In recent times, the WWW has changed the way people interact with one another, express their opinions, provide feedback on websites, and share feelings on social media on different platforms. In the age of AI, it becomes necessary to understand the sentiments from peoples’ opinions, feedback, and interactions. Sentiment analysis has recently drawn a lot of attention. Sentiment analysis has been extensively used and applied in a range of domains, including business and e-commerce. Mainly in the e-commerce sector, where product reviews and service reviews are one of the most important criteria for improving products and services, so analyzing the sentiments behind customer reviews or feedback becomes crucial. On the other hand, sentiment analysis helps customers in decision-making. The authors explored the Amazon reviews dataset, using state-of-the-art natural language-based transformer models to analyze product review sentiment. This paper presents a comparative investigation of different techniques used for sentiment analysis in product reviews to discover which AI-based technique works pre-eminent for review datasets.