| dc.description.abstract | Stunting is a chronic nutritional problem that remains a major challenge in public health development in Indonesia, including in Medan City, as it directly affects the quality of human resources in the long term. This condition is influenced by various multidimensional factors such as socioeconomic conditions, environmental sanitation, access to clean drinking water, parenting patterns, and the availability of healthcare services. This study aims to map regional stunting risk in Medan City by integrating the Clustering Using Representatives (CURE) method for regional grouping and SHapley Additive exPlanations (SHAP) for identifying the most influential risk factors. The dataset used in this study consists of 16,525 toddler records obtained from the Medan City Health Office and has undergone data preprocessing. Six main features were selected for the clustering process based on data completeness and relevance to stunting risk, namely exclusive breastfeeding, income level, health insurance ownership, housing condition, sanitation, and drinking water sources. The clustering results produced four regional clusters with excellent clustering quality, as indicated by a Silhouette Score of 0.899 and a Davies–Bouldin Index of 0.131, confirming that the clusters are compact and well separated. The SHAP analysis reveals that parental occupation (0.6523), sanitation (0.2507), and access to clean drinking water (0.2316) are the most dominant factors influencing stunting risk, followed by health insurance ownership (0.2205), exclusive breastfeeding status (0.1623), income level (0.1603), and housing condition (0.0803). The integration of CURE and SHAP provides a stunting risk mapping that is not only descriptive in illustrating the spatial distribution of risk but also explanatory in identifying the main contributing factors, thereby serving as a strong foundation for formulating targeted and sustainable region-based stunting intervention policies. | en_US |
