Analisis Cluster pada Kabupaten/Kota dengan Algoritma K-Means Berdasarkan Indikator Gizi Buruk Balita di Provinsi Sumatera Utara Tahun 2023

Abstract

One of the biggest challenges in global health, especially in developing countries, is malnutrition in toddlers. This nutritional problem is not just a statistical number, but has a direct impact on the growth and development of children in the future. This study aims to (1) Apply the cluster K-means algorithm method to group districts/cities based on malnutrition indicators of toddlers, (2) Determine the optimal number of clusters in data grouping, (3) Determine the variables that have the greatest difference in the clusters formed. There are four variables measured for this grouping, namely Low Birth Weight (BBLR), Exclusive Breastfeeding, Posyandu, and Poverty. In this study, the data used are secondary data from the Sumatra Provincial Health Office and the Central Statistics Agency (BPS) of North Sumatra Province. Data processing is carried out with the help of R and SPSS software. From the results of the research of the K-means algorithm with the help of R software in determining the number of clusters, 2 clusters were obtained. Cluster 1 with members of the regency/city group with high levels of malnutrition which contains 3 regencies/cities, namely Deli Serdang Regency, Lalat Regency, and Medan City. Cluster 2 with members of the regency/city group with low levels of malnutrition consists of: Nias Regency, Mandailing Natal, South Tapanuli, Central Tapanuli, North Tapanuli, Toba, Labuhan Batu, Asahan, Simalungun, Dairi, Karo, South Nias, Humbang Hasundutan, Pakpak Bharat, Samosir, Serdang Bedagai, Batu Bara, North Padang Lawas, South Labuhan Batu, North Nias, West Nias, Sibolga City, Tanjung Balai City, Pematang Siantar City, Tebing Tinggi City, Binjai City, Padang Sidempuan City, Gunung Sitoli City. Based on the significance value of each variable with a value less than 0.05 in the ANOVA table, so that the variable that provides the greatest difference in the two clusters formed is the posyandu variable, which is with an F value of 125,739 with a significant value of 0.000.