Using Machine Learning to Predict the Defaults of Credit Card Clients

With the rapid development of the Internet, the convenience of electronic transfer and the rapid expansion of credit card business, the use of credit cards in daily life is becoming very popular. However, credit card debt is a significant threat to financial organizations, government institutions, and even individual investors. Therefore, methods to predict credit cards’ risk effectively, quickly, and accurately have become increasingly necessary. In this research, we propose a framework that makes use of simple feedforward neural network or artificial neural network (ANN), together with a binary feature selection algorithm to enhance the accuracy of the network. The simulation results show that our approach is highly competitive compared to other methods.