Predictive Analytics for Credit Risk Assessment: A Machine Learning Approach to Loan Default Prediction

The integration of Machine Learning (ML) into financial decision-making helps the financial to strengthen credit-risk assessment and improve lending decisions. Traditional credit assessment methods that depend on fixed criteria and historical records, often struggle to capture complex relationships among borrower and loan characteristics. Machine learning helps bridge that gap. This study examined the application of Machine Learning techniques to loan default prediction. Using a dataset comprising 45,000 loan records, the study analysed borrower demographic, financial, employment, credit-history and loan-related characteristics to identify factors associated with default risk. The study applies three supervised machine learning classification models namely Logistic Regression, Decision Tree, and Random Forest and compared results using accuracy, precision, recall, F1-score and ROC-AUC. Random Forest performed best overall (92.69% accuracy, 88.19% precision, 81.51% F1-score, 97.04% ROC-AUC), while Decision Tree achieved the highest recall (77.46% vs. 75.77%), making it slightly better at identifying actual defaulters. Logistic Regression also performed well, reaching 94.95% ROC-AUC. Random Forest offers the strongest overall performance, but higher-recall models are valuable when minimizing missed high-risk borrowers is the priority. Overall, Machine Learning proved a practical tool for strengthening credit-risk assessment and supporting better financial decisions.