PERBANDINGAN KINERJA ALGORITMA K-MEANS DAN K-MEDOIDS UNTUK MENGELOMPOKAN DATA SISWA PENERIMA BOSDA DI SMKN1 KATIBUNG

Abstract

Regional Operational Assistance (Bosda) for SMKN 1 Katibung is assistance provided to students who cannot afford it in the form of assistance with Educational Development Contribution (SPP) costs for 1 year.

Taking Regional School Operational Assistance (BOSDA) at SMKN 1 Katibung still uses manual selection, namely involving several crucial stages. The author is interested in conducting a comparative analysis of the K-means and K-Medoids clustering algorithms. Based on the research results, it can be concluded that the modeling was carried out using using the K-means algorithm does not produce good results. The results of clustering using the K-means algorithm show a Davies Bouldin Index (DBI) value of 0.842, which indicates that the resulting data partition is not optimal enough. However, by using the K-Medoids algorithm, the clustering results show a significant improvement in the quality of the data partition. The DBI value is 0.671. The increase in clustering quality, the results of the research, show an increase of around 20.33%, indicating that clustering carried out using the K-Medoids algorithm produces better data partitioning than using K-means. The resulting clusters are more distinct from each other and more internally cohesive, indicating that the K-Medoids algorithm is more effective in handling the data and dividing it into better groups. Therefore, in the context of this research, it can be concluded that the use of the K-Medoids algorithm is more recommended than K-means for clustering the same data.