<![CDATA[Regional disparities in welfare remain a significant challenge in Indonesia, particularly in West Java Province, where socioeconomic conditions vary across districts and municipalities. This study aims to classify districts and municipalities in West Java based on socioeconomic characteristics using the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. The variables analyzed consist of Mean Years of Schooling, Number of Poor Population, Poverty Severity Index, Annual Population Growth Rate, and Human Development Index (HDI). Secondary data were obtained from the West Java Provincial Statistics Office (BPS). Prior to clustering, the variables were standardized using the Z-score method and analyzed using Principal Component Analysis (PCA) to reduce dimensionality and address multicollinearity. The first principal component (PC1), which explained approximately 82% of the total variance, was used as the input for DBSCAN clustering. The optimal DBSCAN configuration was determined by evaluating combinations of epsilon (ε) and minimum number of points (MinPts) using the Silhouette Coefficient. The results showed that ε = 0.06 and MinPts = 2 produced the highest Silhouette Coefficient of 0.8879592, resulting in two clusters and three observations classified as noise. Cluster 1 consisted of 22 districts and municipalities and was characterized by a relatively higher average number of poor population and Poverty Severity Index. Cluster 2, consisting of Sukabumi City and Cimahi City, exhibited higher Mean Years of Schooling and HDI, along with a lower Poverty Severity Index. Pangandaran, Cirebon City, and Banjar City were identified as noise due to their distinctive characteristics. These findings demonstrate that DBSCAN can identify heterogeneous socioeconomic patterns among regions in West Java]]>
Copyrights © 2026
