Evaluating the Effectiveness of mHealth Applications for Diabetes Management Using Machine Learning

The rapid increase in diabetes prevalence worldwide demands innovative and technology-driven management solutions. Mobile health (mHealth) applications have emerged as effective tools enabling patients to monitor glucose levels, manage lifestyle habits, and improve treatment adherence. This study evaluates the effectiveness of mHealth applications for diabetes management using machine learning techniques. Multiple factors such as user engagement, clinical outcomes, usability, and behavioural patterns are analysed using Sequential Graph Collaborative Filtering (SGCF), Extreme Gradient Boosting (XGBoost), and Linear Support Vector Classifier (Linear SVC). The proposed framework integrates heterogeneous health and behavioural datasets to identify patterns influencing diabetes outcomes. Experimental results demonstrate that the proposed method achieves 97.5% accuracy, outperforming Autoencoder (86.3%) and SVM (92.2%). The findings highlight the potential of machine learning–enabled mHealth systems to provide personalized healthcare recommendations and improve long-term diabetes management outcomes.