An Investigation of Credit Card Default Prediction Using Machine Learning Classifiers - Decision Tree and SVM

Due to the global financial crisis and elevated credit risk, default forecasting is essential for all economic sectors. Advanced machine learning techniques have replaced traditional linear models for credit default prediction. Big data risk control algorithms now outperform traditional banking techniques in terms of scalability, speed, and accuracy. Support Vector Machine (SVM) and Decision Tree models are compared in this study utilising the German Credit Dataset, which has 21 features and 1000 cases. Following pre-processing that included outlier treatment and category encoding, SVM outperformed Decision Tree with an accuracy of 80.7% versus 72.6%. Through scalable machine learning solutions, these discoveries allow financial institutions to assist small and medium-sized businesses that were previously underserved by traditional banking.