CREDIT CARD DEFAULT PREDICTION USING MACHINE LEARNING: A LOGISTIC REGRESSION MODEL APPROACH

This comprehensive analysis explores the application of logistic regression models in predicting credit card defaults, a critical concern for financial institutions worldwide. Current research demonstrates that logistic regression provides an effective framework for identifying potential defaulters based on demographic, financial, and behavioral features. Studies across various datasets reveal prediction accuracies ranging from 85- 90%, with key predictors including payment history, credit utilization, income levels, and demographic factors. While more complex algorithms may offer marginal performance improvements, logistic regression remains valuable for its interpretability and practical implementation advantages in financial risk management systems.