Fanny Ramadhani, Said Iskandar Al-Idrus, Dian Septiana, Arnita, Diah Retno Wahyuningrum, Salamah
Stunting remains a persistent public health challenge in Indonesia, particularly in North Sumatra, where regional variances prevent effective intervention. This study proposes an integrated machine learning framework to predict and spatially map stunting risk by utilizing multidimensional data sources, including socioeconomic indicators, environmental factors, and health infrastructure data. Employing both supervised learning (Random Forest) and unsupervised learning (K-Means clustering), the framework enables reliable risk classification and geographical segmentation. The Random Forest model produced outstanding classification across four risk categories, while the K-Means clustering yielded coherent clusters verified by a silhouette score of 0.4058. The spatial visualization component blends clustering results with Geographic Information Systems (GIS), giving actionable risk maps associated with the strategic objectives of the North Sumatra Regional Medium-Term Development Plan (RPJMD) 2019-2023. This technique enhances the precision of policy targeting in resource-constrained environments and provides a reproducible model for public health decision support. The findings show the practicality and policy importance of mixing data science with regional planning to reduce stunting through evidence-based therapies. © 2025 IEEE.
Universitas Negeri Medan, Department of Computer Science, Medan, Indonesia; Universitas Negeri Medan, Department of Statistics, Medan, Indonesia; Universitas Negeri Medan, Department of Medan, Indonesia; Universitas Malikussaleh, Department of Diploma Keperawatan, Sigli, Indonesia
