<![CDATA[Abstract Unemployment is a short-term economic problem that significantly affects social conditions and public welfare. The Open Unemployment Rate (OUR) serves as an important indicator to measure unemployment levels. In Indonesia, OUR data are periodically published by the Central Bureau of Statistics (Badan Pusat Statistik, BPS) by province and year. Analyzing these data helps identify regions with high or low unemployment rates to support effective labor policy formulation. This study applies the K-Means clustering algorithm to classify Indonesia’s open unemployment rate using four distance metrics: Euclidean, Manhattan, Chebyshev, and Cosine Similarity. The clustering performance was evaluated using the Davies-Bouldin Index (DBI) to determine the best distance metric. The results indicate that Chebyshev Distance produced the best cluster quality with a DBI value of 0.606, while Cosine Similarity yielded the poorest result with a DBI of 1.445. Therefore, Chebyshev Distance is recommended as the most suitable distance metric for clustering Indonesia’s open unemployment rate using the K-Means algorithm. Keywords: open unemployment, K-Means, distance metric, Davies-Bouldin Index.]]>
Copyrights © 2025
