Pengelompokan Mahasiswa Calon Penerima Bantuan Pendidikan UPZ UNP Menggunakan Metode Simple K-Medoids Partitioning

Abstract

UPZ UNP is a zakat management institution established by BAZNAS of West Sumatra within Universitas Negeri Padang. The current manual selection process for educational assistance recipients is considered inefficient due to the large number of applicants. This study applies the Simple K-Medoids Partitioning method with Gower distance to objectively cluster student data. The data used were obtained from UPZ UNP for the period of July–December 2024. Several cluster numbers (k=2 to k=7) were tested, and validation using the Medoid Shadow Value indicated that k=2 was the most optimal, with a value of 0.463247. Object 45 and object 146 served as the cluster centers. The final result showed that cluster 1 consisted of 405 students and cluster 2 included 265 students. Cluster 2 is prioritized as the recipient group. The findings of this study are expected to support more accurate and efficient decision-making at UPZ UNP.