Penerapan Algoritma K-Means dalam Klasterisasi Wilayah Prioritas Penanganan BBLR dan Risiko Stunting

Abstract

This study aims to classify districts and municipalities in South Sumatra Province based on the risk level of Low Birth Weight (LBW) and severe malnutrition using the K‑Means Clustering algorithm as a basis for mapping priority areas for stunting prevention. The research utilizes secondary data from 2024 consisting of total live births, LBW cases, and severe malnutrition cases across 17 regions. Both risk indicators were transformed into rate-based measurements to ensure proportional comparisons between regions and subsequently normalized using the Min–Max method to equalize variable scales for Euclidean distance computation within the clustering process. The optimal number of clusters was determined through the elbow method combined with the Davies–Bouldin Index (DBI), which indicated that k = 3 provides the most suitable cluster structure for the dataset. The clustering results formed three distinct groups representing low-risk, medium-risk, and high-risk areas. Regions classified into the high‑risk cluster exhibited the highest LBW and malnutrition rates, thus becoming the primary targets for intervention. The findings demonstrate that the K‑Means algorithm is effective for health‑risk mapping using numerical epidemiological data and can serve as a reliable analytical tool to support evidence‑based decision‑making in stunting reduction programs