<![CDATA[Digital transformation in the healthcare sector has driven the development of the SatuSehat application as an integrated health service platform in Indonesia. However, the increasing number of users is accompanied by diverse user reviews that reflect user experiences and satisfaction, requiring systematic analysis to understand these perceptions. This study aims to analyze user sentiment in SatuSehat application reviews using a transformer-based deep learning model, namely IndoBERT. The research method employed is Knowledge Discovery from Data (KDD), which includes data collection through web scraping from the Google Play Store, preprocessing, text transformation, classification, and model evaluation. The dataset consists of 3,000 reviews categorized into three sentiment classes: negative, neutral, and positive. The results indicate that the IndoBERT model achieves an accuracy of 86.56%, with precision of 86.01%, recall of 86.56%, and an F1-score of 84.33%. The findings reveal that negative sentiment dominates, primarily related to technical issues such as system errors and login difficulties. In conclusion, IndoBERT is effective in understanding the context of the Indonesian language in informal reviews and can serve as a foundation for developing automated sentiment monitoring systems to improve digital healthcare service quality]]>
Copyrights © 2026
