<![CDATA[In 2024, the number of internet users in Indonesia reached 221.56 million, accounting for 79.5% of the population an increase of 1.4% from the previous year (APJII). This growth has driven digital transformation in various sectors, including healthcare. To support this, the government launched the Mobile JKN app as part of the digitalization of the National Health Insurance (JKN) program, aimed at expanding access to services, especially in remote areas. Despite over 50 million downloads, the app still faces technical issues such as difficulties with registration, verification, and frequent updates that disrupt user experience. This study analyzes user complaints using sentiment analysis with the Support Vector Machine (SVM) algorithm and topic modeling via Latent Dirichlet Allocation (LDA). A total of 285,661 reviews from the Google Play Store (June 2016–December 2024) were collected and pre-processed. Of these, 181,657 reviews were analyzed—80% used for training (145,615) and 20% for testing (36,042). The SVM model showed strong performance, achieving 90% accuracy, 90% precision, 89% recall, and an F1-score of 89%. It classified 12,965 reviews as positive and 23,077 as negative. Topic modeling of negative reviews revealed five key themes with a coherence score of 0.5064: app usage, login and registration, data verification, online services and data changes, and app updates. Further analysis of version 4.12.0 informed improvement recommendations, particularly regarding phone number verification, login, and facial recognition issues.]]>
Copyrights © 2025
