Spatial Clustering of Child Malnutrition in Central Java: A Comparative Analysis Using K-Means and DBSCAN

The issue of malnutrition in children poses a serious challenge in the effort to achieve the well-being and development of a smart generation of children. Central Java is a province on the island of Java with the highest prevalence of stunting, so efforts for improvement and health intervention planning need to focus on areas with malnutrition among children in Central Java. This research aims to identify the spatial patterns and distribution of stunted children in the 35 districts and cities in Central Java using clustering techniques. The data used includes the nutritional status of children in all districts and cities in Central Java. Two clustering methods, K-Means and DBSCAN, were applied to identify groups of districts/cities with stunting characteristics. K-Means resulted in three clusters: low stunting prevalence (11 districts/cities), moderate (18 districts/cities), and high (6 district/cities). DBSCAN grouped 21 districts/cities into one main cluster and identified 14 other districts/cities as outliers. In this study, K-Means outperformed DBSCAN, with a higher Silhouette score (0.403) and a lower Davies-Bouldin Index (0.785).