Perbandingan K-Means dengan Hierarchical Clustering Untuk Pengelompokkan Tingkat Pengangguran di Sumatera Utara

Abstract

Unemployment is a crucial issue faced by many countries, including Indonesia. This study compares two data clustering methods, K-Means and Hierarchical Clustering, to group districts/cities in North Sumatra based on the Open Unemployment Rate (OUR). The K-Means method is known for its speed and simplicity in partitioning data into clusters by determining centroids as the central points, while Hierarchical Clustering organizes data into a more complex hierarchy without requiring a predefined number of clusters. The OUR dataset used in this study was obtained from various districts/cities in North Sumatra and processed using statistical software to apply both methods. The results indicate that the K-Means method provides superior clustering quality with a Silhouette Score of 56.50%, compared to Hierarchical Clustering, which obtained a score of 43.69%. These findings suggest that the K-Means method is more effective in identifying unemployment patterns in the region. This insight can serve as a reference for policymakers in formulating more targeted strategies to address unemployment in North Sumatera.