Exploration of Stopword Removal Application in Ecommerce Product Review Classification Using Machine Learning Model

Online shopping has become a new way of shopping since the pandemic began, people started buying goods online instead of having to go to offline stores to buy everyday goods. E-commerce as a platform used for this online store displays products through images and videos of the goods being sold, ratings and reviews are parameters used by customers to decide which store they will buy their goods from. With thousands of users, there are many reviews from various impressions made by these users that are displayed on the platform. If these reviews are not classified, they can confuse buyers who need reviews before buying a product. The development of a Machine Learning model can be done to help classify text data containing E-Commerce user reviews. However, to get optimal results, the data used and the data cleaning process must be considered. This study discusses the classification of e-commerce product reviews, comparing Naive Bayes and Support Vector Machine (SVM) and the effect of implementing one of the stages in data preprocessing, namely stopword removal. The SVM model without Stopword Removal provides better performance in sentiment analysis with an accuracy of up to 99.9 %.