Perbandingan Metode K-Means, K-Medoids, dan Fuzzy C-Means dalam Pengelompokan Kota dan Kabupaten di Provinsi Sumatera Utara Berdasarkan Indikator Indeks Pembangunan Manusia Tahun 2023

Abstract

In March 2023, the percentage of poor people in urban areas was 8.23% and in rural areas was 8.03%, with a decrease of 0.40 points in urban areas and an increase of 0.07 points in rural areas compared to September 2022. Based on these facts, it can be concluded that there has not been even development in the fields of health, education and the economy between villages and city. Grouping cities/regencies based on Human Development Index (HDI) indicators is one solution for identifying areas that require more attention in efforts to improve community welfare. Cluster analysis is a tool that can be used to group cities and regencies in North Sumatra Province. This research analyzes the comparison of the K-Means, K-Medoids and Fuzzy C-Means methods in conducting cluster analysis. Where in the K-Means method, cluster 1 has 4 cities/regencies, cluster 2 has 8 cities/regencies and cluster 3 has 21 cities/regencies. In the K-Medoids method, the cluster results are the same as the cluster results of the K-Means method. Meanwhile, in the Fuzzy C-Means method, Cluster 1 consists of 9 cities/regencies, cluster 2 consists of 16 cities/regencies and cluster 3 consists of 8 cities/regencies. Then, based on the Davies Bouldin Index (DBI) values from the three methods, it was concluded that the cluster results of the K-Means method were more accurate compared to the K-Medoids and Fuzzy C-Means methods because they had the smallest DBI value.