<![CDATA[This study compares four clustering techniques AHC-Ward, K-Means, GMM, and BIRCH to group sub-districts in North Aceh Regency based on plantation production data. The goal is to find the most effective clustering method for identifying diverse production patterns. Each algorithm's performance was evaluated using three internal metrics: Silhouette, Davies-Bouldin Index (DBI), and Dunn Index. The analysis showed that AHC-Ward had the most consistent and best performance. This method achieved the highest Silhouette score (0.515), the lowest DBI (0.309), and the highest Dunn Index (0.816). In contrast, K-Means performed the worst, while GMM and BIRCH were in between. Based on the optimal AHC-Ward clustering results, three clusters were found that reflect geographical specialization: a cluster with diverse production, a cluster dominated by oil palm, and a cluster focused on Arabica coffee and tobacco. These findings provide crucial insights for local governments to formulate agricultural policies that are more focused and relevant to the specific characteristics of each cluster.]]>
Copyrights © 2025
