IMPLEMENTASI ALGORITMA CLUSTERING DENGAN MODIFIKASI METODE ELBOW UNTUK MENDUKUNG STRATEGI PEMERATAAN BANTUAN SOSIAL DI KABUPATEN BOJONEGORO

Abstract

This research aims to obtain clustering results from the K-Means and K-Medoids methods for grouping communities per sub-district in Bojonegoro Regency based on several types of Social Assistance. The method used is one of the data mining methods, namely the clustering method. As a comparison, the clustering methods used are the K-Means and K-Medoids methods. The research variable used is annual data on Family Social Assistance from the Bojonegoro Regency Social Service from 2020 to 2021, which contains several types of assistance, namely Regional Contribution Assistance, Regional Direct Assistance, Village Direct Assistance, Social Food Assistance Affected by the COVID-19 Pandemic, and the Family Hope Program in the form of discrete data with a ratio scale. The results obtained by modifying the Elbow method obtained the best clustering method is the K-Means method with 5 optimal clusters. The sub-district clusters in 2020 and 2021 are different because in 2020 there is more focus on COVID-19 pandemic assistance, while in 2021, there is more focus on direct Regional and Village assistance