<![CDATA[Occupational risk plays a crucial role in employment insurance management due to its direct impact on claim frequency and the sustainability of insurance systems. However, workers’ risk levels are heterogeneous and influenced by factors such as age, wages, working hours, and claim costs. This study aims to determine the optimal number of clusters and analyze worker risk segmentation using the K-Medoids Clustering method. Secondary data, obtained from Kaggle and consisting of 54,000 observations, were used. From this dataset, 300 samples were selected using the Slovin formula. The analyzed variables included age, weekly wages, hours worked per week, initial claim cost, and final claim cost. Prior to clustering, the data were standardized using a robust scaler and tested for multicollinearity. The optimal number of clusters was determined using the Silhouette Coefficient method. The results indicated that the optimal clustering structure consisted of three clusters, with a Silhouette Coefficient value of 0.703. These clusters represented low-risk, medium-risk, and high-risk worker groups. The findings offer valuable insights for insurers to enhance risk segmentation, claims management, and more targeted premium policy formulation.]]>
Copyrights © 2026
