Early Detection Childhood Diabetes Risk Using Random Forest on Health Data to Support School Meal Policies in Indonesia

The increasing prevalence of childhood diabetes poses a significant public health challenge globally, particularly in developing countries like Indonesia. Early detection and targeted interventions are essential to mitigate the long-term health impacts of diabetes in children. This paper explores the use of machine learning, specifically the Random Forest algorithm, to predict the risk of childhood diabetes using health data, aiming to support school meal policies in Indonesia. The proposed solution leverages a Random Forest model to analyze various health indicators, such as BMI, family history, and lifestyle factors, to identify children at higher risk of developing diabetes. This study contributes to the field by demonstrating the applicability and effectiveness of predictive analytics in public health policy, especially in educational settings. The model's performance is validated using a to-fold cross-validation technique, achieving high accuracy and robust predictive capability. The findings suggest that integrating Random Forest models into school health programs can provide actionable insights for policymakers to design targeted nutritional interventions, thereby reducing the incidence of childhood diabetes and improving overall child health outcomes in Indonesia.