Abstrak
Unemployment is one of the pressing issues faced by by Indonesia as a developing country. According to data from Badan Pusat Statistik (BPS), unemployment is a significant problem in West Java Province, which consistently reports an unemployment rate above the nasional average compared to other provinces in the country. Therefore, addressing this issue in West Jawa Province is critical. To tackle this challenge, an analysis of districts and cities within West Java Province is critical. To tackle this chalengge, an analysis of districts and cities within West Java is necessary to identify areas with high unemployment rates by applying clustering techniques. Clustring encompasses a variety of methods, particularly within the realms of supervised and unsupervised approaches. One prominent unsupervised clustering technique is K-Means. This method is widely used in addressing social issues due to its simplicity, quick convergence, computational efficiency, and ease of implementation. This study aims to employ Z-Score standardization prior to applying the K-Means algorithm for clustering and to evaluate the impact of Z-Score standardization on the clustering process itself. In this research, experiments were conducted by testing the hyperparameter k ranging from 2 to 10 using Silhouette Score for evaluation. The optimal result was achieved with k = 3, yielding a score of 4.305. The application of Z-Score standardization played a significant role in preventing data dominance and ensuring unbiased clustering outcomes, as it normalized the varying scales of the variables.
