Comparison of Sentiment Analysis Model for Shopee Comments on Google Play Store

Abstract

The current COVID-19 pandemic has greatly changed the order of consumption and the Indonesian economy. During the health crisis that hit Indonesia, the e-commerce sector experienced very rapid development because of changes in consumer behavior that are looking for safe and comfortable shopping alternatives. During the COVID-19 pandemic, Shopee became the number 1 online shopping site in Indonesia. However, this cannot be used as a standard for user satisfaction. User satisfaction can only be measured from comments by Shopee application users through the comments and rating features provided by the Google Play Store. Therefore, to be able to find out public opinion about Shopee, a sentiment analysis of the Shopee application will be carried out which can later be used by management to develop even better applications. In this study, the dataset taken is the rating and reviews of Shopee application users on the Google Play Store using the Multinomial Naïve Bayes method, Random Forest Classifier, Logistic Regression, Support Vector Machine, K-Nearest Neighbors, and Extra Trees Classifier. This study uses 1000 comment and rating data which are processed using the Python language. The results of this study indicate that the method that has the highest level of accuracy is the Support Vector Machine algorithm with an accuracy of 88%, Extra Trees Classifier with an accuracy of 86%, Logistic Regression with an accuracy of 85%, Random Forest Classifier with an accuracy of 85%, K- Nearest Neighbors with an accuracy of 83%, and the last is Multinomial Naïve Bayes with an accuracy of 78%.