Classification of Product Review Sentiment by NLP and Machine Learning

Abstract

Online marketing and e-commerce firms were already prospering in Bangladesh during this era of internet technology. Because people are under lockdown due to the COVID-19 epidemic, internet shopping has become the major platform for purchasing because it is the safest option. It sped up the time it took for firms to go online. More online product service providers improve people's lives, but it also raises concerns about product quality and service. As a result, it is simple for new clients to dupe while purchasing online. Our objective is to create a system that uses Natural Language Processing to assess client feedback from online purchasing and deliver a ratio of good and bad comments written in Bangla from past customers (NLP). We gathered approximately 6000 comments and views on the product to conduct the study. As classification approaches, we used sentiment analysis, as well as KNN, Decision Tree, Support Vector Machine (SVM), Random Forest, and Logistic Regression. With an accuracy of 94.78 percent, SVM outperformed all other methods.