<![CDATA[ The loan eligibility assessment for retired employees is a crucial process requiring quick and accurate evaluation to support sound credit decision-making. This research develops a pension loan eligibility classification system using the K-Nearest Neighbors (K-NN) method at Bank BTPN Palembang. The dataset consists of 240 retired customers with attributes including Pension Salary, Loan Amount, 70% Salary, Monthly Installment, Remaining Salary, Credit History, Age, and Target Label. Data preprocessing includes duplicate removal, normalization, encoding, and train-test splitting. The optimal k value is determined using the Elbow method. Experimental results show that K-NN with k = 5 achieves an accuracy of 95.83%. Precision values reach 0.92 for the Eligible class and 1.00 for the Ineligible class, while recall scores are 1.00 and 0.92, with F1-scores of 0.96 for both classes. The confusion matrix indicates 45 correct classifications out of 48 test cases, with only 3 misclassifications. These results demonstrate that the K-NN method effectively supports objective loan eligibility assessments in the financial sector.]]>
Copyrights © 2025
