Perbandingan Metode K-Means dan K-Medoids Dengan Validitas Davies-Bouldin Indeks, Dunn Indeks dan Indeks Connectivity Pada Pengelompokkan Masyarakat Penerima Bantuan Langsung Tunai

Kilo, Nur Ain
Katili, Muhammad Rifai
Hasan, Isran K

Research in the Mathematical and Natural Sciences
Vol. 4 No. 1 (2025): November 2024-April 2025

This study discusses the comparison between the K-Means and K-Medoids methods in grouping direct cash assistance (BLT) recipients, with an assessment using three validity indices: Davies-Bouldin Index (DBI), Dunn Index, and Connectivity Index. The main objective of this study is to determine the most effective clustering method for grouping BLT recipient data by considering the quality of the resulting clustering. In the experiment, the K-Means method with three clusters produced, namely: Cluster 1 with 10 family head members, Cluster 2 with 101 family head members, and Cluster 3 with 118 family head members. In contrast, the K-Medoids method also with three clusters, namely: Cluster 1 with 67 family head members, Cluster 2 with 59 family head members, and Cluster 3 with 103 family head members. Based on the evaluation using the Davies-Bouldin Index and Connectivity Index, the K-Means method showed better performance than K-Medoids. The DBI value for the K-Means method is 1,307, while the Connectivity Index value is 40,079, which shows that the K-Means clustering results are more effective in producing separate and quality clusters in the context of grouping BLT recipient communities.