Grouping of Regencies/Cities in South Sulawesi Based on Horticultural Production Using Cluster Analysis

Abstract

In this research we use Cluster Analysis which is a multiple variable technique which in this research has the main aim of grouping districts/cities and comparing hierarchical clustering analysis groupings using the Single Linkage Method, Complete Linkage, Average Linkage, Centroid Method and Ward Method based on horticultural production in South Sulawesi. By using Euclidean Distance, the clustering in the Single Linkage Method, Average Linkage and Centroid Method has 4 clusters, while in the Complete Linkage Method and Ward Method there are 5 clusters. Based on the clusters, we will see a comparison using the CTM (Cluster Tightness Measure) formula in the Single Linkage Method, Average Linkage, Centroid Method which produces the same value, namely 0.102248, while the Complete Linkage Method and Ward Method produce a value of 0.105316. It can be concluded that the best method for clustering is Single Linkage, Average Linkage and Centroid Method.