Sentiment Analysis of User Reviews of GoPay Services on the Gojek Application Using the Naïve Bayes Method

Abstract

The development of digital technology has driven the emergence of various application-based financial services, including GoPay within the Gojek platform. GoPay facilitates cashless transactions such as payments for transportation, food, and bills. As the number of users increases, various reviews reflecting satisfaction and complaints about the service have emerged. This study aims to analyze user sentiment toward the GoPay service in the Gojek application using the Naïve Bayes method. User review data were obtained through a web scraping process totaling 1,000 entries, followed by data cleaning, text normalization, tokenizing, and sentiment labeling into positive, negative, and neutral categories. The TF-IDF method was used for word weighting to convert text into numerical form before classification with the Multinomial Naïve Bayes algorithm. The results show that the Naïve Bayes model is capable of classifying user reviews with an accuracy rate of 90%. Based on the confusion matrix and classification report, the neutral sentiment category dominates with the highest recall value of 1.00, while positive and negative sentiments have lower proportions. This indicates that most users provide neutral feedback regarding the GoPay service in the Gojek application. Overall, the Naïve Bayes method has proven effective in analyzing user sentiment toward digital services such as GoPay, making the results of this study a useful reference for Gojek developers in improving service quality and user satisfaction.