<![CDATA[Stunting is a chronic nutritional problem that remains a national priority issue in Indonesia. According to the 2022 Indonesian Nutrition Status Survey (SSGI), the national stunting prevalence reached 21.6%, with a target reduction to 14% by 2024. Accurate prediction of stunting risk remains a challenge, particularly in regions like Palembang City, which exhibit diverse socio-economic conditions and complex anthropometric characteristics. This study develops a hybrid machine learning model for stunting risk prediction by integrating classification algorithms with a Genetic Algorithm (GA) for feature selection. The hybrid approach aims to enhance predictive accuracy and efficiency based on numerical and socio-economic data. A total of 6,000 samples were used, and after preprocessing (trimming, winsorization, normalization, and SMOTE), 5,366 clean data samples were obtained. Four classification algorithms were tested: Decision Tree, K-Nearest Neighbor, Random Forest, and XGBoost. The best performance was achieved by the XGBoost model, with an accuracy of 84.08%, recall of 93%, and F1-score of 0.91 for the majority class. By integrating the Genetic Algorithm, optimal accuracy reached 95.34% in the third generation of feature selection. This study contributes a hybrid machine learning-based predictive framework that can be adopted by local health institutions for more targeted early detection of stunting risk.]]>
Copyrights © 2025
