<![CDATA[The increasing adoption of mobile banking has generated a large volume of user reviews on the Google Play Store, providing valuable insights into customer satisfaction with digital banking services. This study analyzes user sentiment toward four mobile banking applications BRimo, Livin' by Mandiri, BCA Mobile, and SeaBank—and compares the performance of the Naïve Bayes and Support Vector Machine algorithms for sentiment classification. A dataset of 40,000 Play Store reviews was collected through web scraping. Reviews were labeled based on user ratings, followed by preprocessing, TF-IDF feature extraction, and sentiment classification. Model performance was evaluated using the train-test split method and 5-fold cross-validation. The results indicate that Naïve Bayes outperformed and demonstrated greater stability than SVM across all evaluation metrics, achieving an accuracy of 89.23%, precision of 89.76%, recall of 89.23%, and F1-score of 89.37%, while SVM achieved an accuracy of 88.39%. Sentiment analysis revealed that SeaBank received the highest number of positive reviews, whereas BCA Mobile recorded the highest proportion of negative sentiment. These findings provide a comparative evaluation of Naïve Bayes and SVM for mobile banking sentiment classification and offer practical guidance for selecting appropriate classification algorithms while supporting the continuous improvement of digital banking services through user feedback.]]>
Copyrights © 2026
