International Journal of Science, Engineering and TechnologyAuthors: Mrs. Narni Sree Ratna Niharika, Mr. V Anil Santosh
Abstract: Banks serve as the foundation of the global financial system, generating considerable income from loan interest. Nevertheless, loan defaults can turn expected profits into substantial losses, highlighting the necessity for thorough risk evaluation before loan sanctioning. In this research, we utilized advanced machine learning methods to accurately and efficiently predict loan default risk. Six sophisticated models Decision Tree, Random Forest, Support Vector Machine (SVM), Multi-layer Perceptron (MLP) Neural Network, Naive Bayes, and an innovative stacking ensemble were trained on an extensive dataset consisting of twenty vital attributes from loan applications. The stacking ensemble model attained the highest performance, achieving an accuracy of 78.75%, while the Random Forest model exhibited similar effectiveness at 78.15% with greater computational efficiency. Our examination pinpointed key indicators of credit risk, including loan amount, checking account status, customer age, loan duration, and loan purpose. These results reinforce the transformative capacity of machine learning in advancing credit risk assessment, providing banks with actionable insights for prudent lending choices.
DOI: http://doi.org/10.5281/zenodo.16900081
