Loan Default Prediction Model Improvement through Comprehensive Preprocessing and Features Selection

Abstract

For financial institutions and the banking industry, it is very crucial to have predictive models for their financial activities, as they play a major role in risk management. Predicting loan default is one of the critical issues that they focus on, as huge revenue loss could be prevented by predicting customer’s ability to pay back on time. In this paper, different classification methods (Naïve Bayes, Decision Tree, and Random Forest) are being used for prediction, comprehensive different pre-processing techniques are being applied on the dataset, and three different feature extractions algorithms are used to enhance the accuracy and the performance. Results are compared using F1 accuracy measure, and an improvement of over 3% has been obtained.