Comparison Analysis of Hierarchical Clustering and K-Means Methods in Grouping Provinces in Indonesia Based on Dengue Hemorrhagic Fever (DHF) Cases

Indonesia, a tropical country, experiences climate variations that influence the spread of infectious diseases, including Dengue Hemorrhagic Fever (DHF). The increase in DHF cases necessitates clustering provinces based on their vulnerability to design effective mitigation strategies. This study compares two clustering methods: Hierarchical Clustering and K-Means Clustering. Within the hierarchical clustering analysis, five linkage methods were evaluated: Average Linkage, Complete Linkage, Single Linkage, Ward’s Method, and Centroid Linkage. The best linkage method was identified using the cophenetic correlation coefficient, indicating that Average Linkage produced the most representative cluster structure, resulting in three distinct groups. For the K-Means method, the optimal number of clusters was determined using the Silhouette Coefficient, which also indicated three clusters. Clustering performance evaluation revealed that Average Linkage outperformed K-Means, with a higher Silhouette Score of 0.552. The resulting clusters categorized provinces into three risk groups: high-risk areas (e.g., DKI Jakarta), moderate-risk areas (e.g., West Java and East Java), and low-risk areas, comprising the remaining provinces in Indonesia