<![CDATA[Stunting remains a significant public health issue in Indonesia due to its long-term impact on human resource quality and economic productivity. Despite various intervention programs, disparities in stunting prevalence across regions remain high, particularly in areas characterized by diverse socioeconomic conditions. This study aims to identify regional patterns and group areas based on stunting risk levels using two machine learning approaches: Hierarchical Clustering (HC) and Deep Embedded Clustering (DEC). The data used in this study consist of aggregated toddler measurement data, including the number of toddlers measured, the number of stunting cases, and the percentage of stunting during the 2020–2024 period. The analysis was conducted by comparing the clustering results generated by both methods. The HC method was implemented using the Agglomerative Clustering approach with the Ward linkage criterion, while DEC employed a layered autoencoder architecture optimized using Kullback–Leibler divergence. Cluster quality was evaluated using the Silhouette Score metric. The results show that HC achieved the highest Silhouette Score of 0.5430, while DEC achieved 0.4874, with both methods exhibiting year-to-year performance variation. These findings indicate that HC provides better clustering stability, whereas DEC demonstrates greater adaptability to data complexity and nonlinear patterns. The integration of both methods offers a comprehensive big data–driven health analytics framework, representing an innovative approach for evidence-based decision-making in identifying and addressing stunting-prone regions.]]>
Copyrights © 2026
