Analisis Cluster dengan Average Linkage Method dan Ward’s Method pada Pengelompokan Kabupaten/Kota di Provinsi Sumatera Utara Berdasarkan Indikator Indeks Pembangunan Manusia Tahun 2022

Abstract

Cluster analysis is an interdependence technique that can describe the proximity or similarity between objects and variables. Cluster analysis is divided into two methods: hierarchical and non-hierarchical. This study employs hierarchical cluster analysis using the average linkage method and Ward's method. Both methods are applied to cluster the districts/cities in North Sumatra Province based on Human Development Index (HDI) indicators. From the 33 districts/cities in North Sumatra Province, both the average linkage and Ward's methods produced 3 groups: low-level, medium-level, and high-level clusters. The low-level cluster includes 9 districts/cities: Pakpak Bharat, Padang Lawas, Nias Selatan, Nias Barat, Nias Utara, Humbang Hasundutan, Nias, Gunungsitoli City, and Samosir. The medium-level cluster consists of 23 districts/cities: Mandailing Natal, Tanjungbalai City, South Tapanuli, Padang Lawas Utara, Batu Bara, Central Tapanuli, Asahan, South Labuhanbatu, North Tapanuli, Serdang Bedagai, Langkat, Dairi, Sibolga City, Padangsidimpuan City, North Labuhanbatu, Labuhan Batu, Toba, Tebing Tinggi City, Simalungun, Karo, Deli Serdang, Binjai City, and Pematangsiantar City. The high-level cluster consists of a single member, Medan City. In this study, Ward's method proved to be more effective as it achieved a lower Sum of Squared Errors (SSE), indicating that it can form more compact and homogeneous clusters. Additionally, dendrogram evaluation revealed that Ward's method produces clusters that are more uniform and consistent.