^Impact of Machine Learning on Dietary and Exercise Behaviors in Type 2 Diabetes Self-Management: A Systematic Literature Review^ Sukesih, Heny Siswanti, Ridlwan Kamaluddin
Universitas Jenderal Soedirman
Abstract
Abstract
This systematic literature review (SLR) explores the role of machine learning in dietary and exercise behaviors for type 2 diabetes self-management. While prior studies often examine diet and physical activity separately, their combined impact remains underexplored. Using a systematic search across ten digital libraries (2019-2023), 70 primary studies were identified and assessed through structured quality criteria. Results indicate that machine learning provides effective tools for personalized dietary recommendations, physical activity monitoring, and complication prediction. Algorithms applied include supervised, semi-supervised, and unsupervised learning, leveraging both public and non-public datasets on dietary patterns, activity levels, and glucose biomarkers. A key research gap was identified: limited studies integrating diet and exercise simultaneously. Addressing this gap is essential for holistic, personalized diabetes care. Overall, integrating machine learning with lifestyle factors shows strong potential to improve self-management, reduce complication risks, and ease healthcare burdens, offering valuable insights for researchers and practitioners.
Keywords: Type 2 diabetes, self-management, diet, exercise, machine learning
