Comparison of Naïve Bayes and SVM Methods in Sentiment Analysis of User Reviews on the RSUD AL IHSAN Mobile Application

Abstract

The advancement of digital services in the healthcare sector necessitates continuous evaluation of user experience. One such innovation is the RSUD AL IHSAN Mobile application, which provides digital access to hospital information and administrative services. This study aims to analyze user sentiment toward the application available on Google Play Store using two popular text classification algorithms: Naïve Bayes and Support Vector Machine (SVM). As of now, the RSUD AL IHSAN Mobile application has been downloaded over 100,000 times, with a rating of 4.6 and 1,540 user reviews. A total of 1,500 reviews were collected via web scraping for analysis. The collected data underwent preprocessing, sentiment labeling into positive and negative categories, and model training using both algorithms. Evaluation was conducted using accuracy, precision, recall, and F1-score metrics. The results show that the SVM algorithm performed better, achieving 100% accuracy, while Naïve Bayes reached 98.84%. The WordCloud visualization highlights differences in focus between positive and negative reviews, indicating which service aspects are appreciated and which require improvement. These findings are expected to serve as valuable input for application developers in enhancing the quality and user experience of RSUD AL IHSAN Mobile.