<![CDATA[This study compares the performance of K-Means and DBSCAN algorithms in clustering 38 Indonesian provinces based on six infectious disease indicators in 2025, namely Tuberculosis (TB), HIV/AIDS, leprosy, malaria, and Dengue Hemorrhagic Fever (DHF). The uneven distribution of diseases across regions requires accurate mapping for targeted public health interventions. The research stages included handling missing values with median imputation, Z-Score normalization, clustering modeling, and evaluation using Silhouette Score, Davies-Bouldin Index (DBI), and Calinski-Harabasz Index (CHI) metrics. Evaluation results indicate that the DBSCAN algorithm (eps=1.2, MinPts=2) consistently outperformed K-Means (K=6) across all three parameters. DBSCAN achieved a Silhouette Score of 0.6282, DBI of 0.4059, and CHI of 21.8583, whereas K-Means only reached a Silhouette Score of 0.4395, DBI of 0.6349, and CHI of 15.8653. DBSCAN's superiority lies in its ability to isolate extreme outliers, such as malaria cases in Papua and DHF in Bali, as noise. In conclusion, DBSCAN is proven to be more robust and representative for clustering national-scale epidemiological data with right-skewed distributions. The resulting cluster map can be utilized by the government to prioritize specific and targeted health program allocations in each region.]]>
Copyrights © 2026
