PERBANDINGAN METODE HASIL EVALUASI PENGELOMPOKAN PENERIMA BANTUAN SOSIAL TUNAI (BST) BERDASARKAN PARAMETER KELAYAKAN BANTUAN SOSIAL MENGGUNAKAN METODE K-MEANS DAN K-MEDOIDS (STUDI KASUS: KECAMATAN SOLEAR)

Abstract

Cash Social Assistance (BST) is a government program aimed at supporting economically vulnerable communities. Accurate identification of eligible beneficiaries is essential to ensure effective and targeted assistance. This study compares the performance of the K-Means and K-Medoids clustering algorithms in grouping BST beneficiaries based on socioeconomic eligibility parameters in Solear District. A quantitative data mining approach was employed using a dataset of 2,000 beneficiary records processed with RapidMiner. The research included data preprocessing, transformation, normalization, clustering, and evaluation using the Davies-Bouldin Index (DBI). Three clusters were generated based on eighteen socioeconomic attributes. The results show that the K-Means algorithm achieved a DBI value of -2.245, outperforming K-Medoids with a DBI value of -1.694. The lower DBI value indicates that K-Means produced more optimal cluster quality and better separation among beneficiary groups. These findings suggest that K-Means is more suitable for supporting objective and data-driven classification of BST beneficiaries.