ABSTRACT
Internet adoption has increased rapidly during the worldwide COVID-19 pandemic. Nowadays people not only prefer to shop using various e-commerce platforms, but also like to provide feedback and express their opinions and experiences using the online platforms. Since new customers try to understand the products' utility and acceptability from other consumers' reviews, it has become crucial to analyze the customers' sentiments and opinions on each product. In this paper, we have presented a sentiment analysis technique on the basis of product reviews written in Bangla language to better understand the combined consumer perspective. Our work aims to compare existing classifiers' performance and find the best algorithm for our dataset. We collected reviews from the leading Bangla bookselling e-commerce site 'Rokomari.com' for this work. We implemented ML and DL classifier models and compared their overall performance on this dataset. The experimental studies show that the best accuracy is achieved from LSTM and SGD over the other implemented ML and DL based classifier models. © 2022 IEEE.