<![CDATA[Diabetes mellitus is a non-communicable disease whose prevalence continues to increase globally and poses a serious challenge to public health systems. The increase in diabetes cases is influenced not only by genetic and clinical factors but also by changes in people's lifestyle behaviors such as unbalanced diets, low physical activity, obesity, and smoking habits. Therefore, a behavioral data-based risk prediction approach is crucial to support early detection and disease prevention strategies at the population level. This study aims to develop a diabetes mellitus risk prediction model based on community lifestyle and physical activity data using a machine learning approach. The study used a quantitative design with a predictive analysis approach. The research data consisted of health behavior variables such as body mass index, physical activity, diet, smoking habits, and sleep duration. The analysis process was carried out through data preprocessing, variable exploration, machine learning model development, and model performance evaluation using accuracy, precision, recall, and ROC-AUC metrics. Several machine learning algorithms were applied to identify patterns of association between behavioral factors and diabetes mellitus risk. The results showed that the developed prediction model was able to identify individuals at risk of diabetes with a good level of accuracy. Body mass index, physical activity level, and dietary patterns emerged as the most influential factors in the prediction model. These findings suggest that lifestyle behavior data can be utilized as important indicators in a community diabetes risk screening system. The resulting prediction model has the potential to support the development of a data-driven early detection system and aid in the formulation of more effective diabetes prevention strategies at the community level.]]>
Copyrights © 2026
