Loan Approval Prediction using Adversarial Training and Data Science

Abstract

Loan approval is critical decision-making in the financial sector, impacting financial stability and reputation. In recent times, many machine learning models have been introduced. However, these models may be biased towards certain groups of borrowers, resulting in unfair loan approval decisions. So, the financial industry requires a fair and accurate prediction model. This paper proposes a model for loan approval prediction that combines Adversarial Training and Data Science techniques. We develop a model by training with a real-time data set, and testing that shows our model achieves better accuracy and fairness than existing models. Our model demonstrates the potential of Adversarial Training and Data Science for improving the Loan Approval Prediction process. This paper contributes to growing research on Adversarial Training and Data Science techniques in the financial sector.