K-Means Clustering Based on Distance Measures: Stunting Prevalence Clustering in South Kalimantan

Stunting interferes with brain growth, intelligence, physical growth, metabolic disorders, decreased cognitive abili-ties, diabetes, obesity, heart disease, cancer, stroke, and immunity. South Kalimantan has the sixth-largest stunting prevalence index out of thirty-eight provinces in Indonesia. This research aims to best classify the prevalence of stunting by comparing the distance measurement method. Researchers used the K-Means clustering method to find the best clusters and measured distances using Camberra, Chebyshev, Euclidean, and Manhattan. After the clusters are formed, the researcher uses the Average Within Centroid Distance and the Davies Bouldin Index to determine the optimal cluster using the elbow method. The results obtained from this study indicate that measuring distances using the Chebyshev method gives three optimal clusters, with cluster 1 having four regencies/cities, cluster 2 having three regencies/cities, and cluster 3 having six regencies/cities. Whereas for distance measurements using Chebyshev, Euclidean, and Manhattan gives the exact five optimal numbers of clusters, where cluster 1 has four regencies/cities, cluster 2 has one regency/city, cluster 3 has six regencies/cities, cluster 4 has one regency/city, and cluster 5 has one regency/city. The clusters obtained can be used as a reference for the government to prioritize the handling of stunting in South Kalimantan.