Analisis Perbandingan Silhouette dengan Elbow pada Algoritma K-Means dan DBSCAN

Khairani Ritonga, Putri
Siddik Hasibuan, Muhammad

METIK JURNAL
Vol. 9 No. 1 (2025): METIK Jurnal

This study aims to apply clustering methods using the K-Means and DBSCAN algorithms to group community data based on parameters such as income, housing condition, occupation, number of dependents, and health status. To determine the optimal number of clusters in the K-Means algorithm, the Elbow and Silhouette methods were employed. The research utilized Python and Google Colaboratory as data analysis tools. The clustering results showed that the DBSCAN algorithm was more effective in identifying homogeneous community groups without the need to predefine the number of clusters, while K-Means produced more structured results but relied on a predetermined cluster count. This research is expected to aid in more accurate and efficient decision-making for community data grouping.