Mobile health app user analysis based on machine learning methods

Abstract

Mobile health and telemedicine are gaining increasing attention from society, with a growing number of mobile health applications being developed and adopted by healthcare professionals and patients. However, there remains a lack of user-focused research on mobile medical applications, particularly in understanding user behaviors within online medical education platforms. This lack of understanding limits our ability to evaluate their effectiveness and guide future development strategies for medical applications.

Therefore, this study examined user behavior records from a mobile health application, which is designed for diabetic patients in Singapore. By reclassifying and analyzing user learning behaviors, anomalous behaviors were identified and filtered out. A user interaction matrix was constructed based on these behaviors, followed by dimensionality reduction using correlation analysis and principal component analysis. Ultimately, users were clustered into six distinct categories, with each group analyzed in detail. The findings provided insights and guidance for the future development and optimization of the application.