Credit Approval Prediction Based on Boruta-GBM Model

In the context of bank loans' application, the choice of high-quality customers has become the banks' key emphasis in work. In this paper, we make statistical analysis on the crucial factors that influence credit approval. Firstly, the Boruta algorithm is utilized to screen the original features. For the new feature combination, Gradient Boosting Machines (GBM) model is used to decide credit approval. Finally, by comparing with other classification models, the Boruta-GBM model greatly improves the accuracy. And because of the simplicity of the Boruta-GBM model, it is convenient to intelligently select high-quality customers for bank loans.