Hierarchical Clustering Kabupaten/Kota di Sumatera Berdasarkan Faktor-faktor yang Berhubungan dengan Indeks Pembangunan Manusia

Abstract

Components of the Human Development Index will be used to cluster regions with similar characteristics and provide information on development conditions in the region, especially in Sumatra. One of the clustering methods that will be used is the centralized hierarchical method using factor scores in principal component analysis to enable clustering. This study aims to examine the clustering of districts/cities in the Sumatra region based on factors related to human development. The results show that the centralized hierarchical clustering in Sumatra based on factors related to human development produces six optimal clusters, except for the centroid method which produces five optimal clusters. The optimal clustering results from the single linkage method, the average linkage method, and the centroid method have their respective similarities. Based on the clustering results for each method, the results of the six clusters of the Ward method provide better clustering performance results seen from the number of cluster members that are more evenly distributed and diverse compared to the other methods.