Prediction of Customer Creditworthiness with the C4.5 Algorithm at PT Menara Indonesia Company

Abstract

Customer credit assessment is still carried out using traditional methods that are time-consuming and less accurate. This is evidenced by the fact that there are still customers with problematic credit who pass the loan application process. To address this issue, this research aims to develop a method using the C4.5 algorithm. The purpose of this research is to improve the efficiency and effectiveness of the company’s debt collection process. This research uses customer data from PT Menara Indonesia, which has credit loans. The data includes seven independent variables: net income, loan amount, credit score, number of arrears, tenure, assets, and loan age, as well as one dependent variable, namely credit risk. The C4.5 algorithm is applied to build a customer credit repayment prediction model. This model is tested using the k-fold cross-validation method with k = 10. The test results show that the C4.5 model has excellent performance, with an accuracy of 99.70%, precision of 99.25%, recall of 98.52%, and an F1-score of 98.88%. The advantage of this method is its ability to provide highly accurate predictions, thereby helping the company identify high-risk customers and improve the overall debt collection process