<![CDATA[The rapid development of digital financial services has led to increased use of digital wallets, one of which is the GoPay application, resulting in a large volume of user reviews. These reviews contain valuable information regarding user satisfaction and service-related issues, making automated methods necessary to accurately analyze user sentiment. This study aims to analyze sentiment in GoPay user reviews and compare the performance of the Naïve Bayes and Support Vector Machine (SVM) algorithms for sentiment classification.This research uses a dataset of 132,393 GoPay user reviews obtained from the Kaggle platform. The data are labeled based on user ratings into three sentiment classes: positive, neutral, and negative. The research stages include text preprocessing, feature transformation using the Term Frequency–Inverse Document Frequency (TF-IDF) method, sentiment classification using the Naïve Bayes and SVM algorithms, and model performance evaluation using accuracy, precision, recall, and F1-score metrics.The results show that 79.2% of the reviews are classified as positive, 17.1% as negative, and 3.7% as neutral. Based on performance evaluation, the SVM algorithm demonstrates superior results with an accuracy of 90.65%, precision of 90.7%, recall of 90.65%, and F1-score of 89.05%, compared to Naïve Bayes, which achieves an accuracy of 87.89%, precision of 89.1%, recall of 87.89%, and F1-score of 88.42%. These findings indicate that SVM is a more optimal method for sentiment analysis of GoPay user reviews, while Naïve Bayes remains an efficient and competitive alternative for large-scale text classification.]]>
