Neural Networks in Finance

8 Classification: Credit Card Default and Bank Failures

This chapter examines how well neural network methods compare with more traditional methods based on discriminant analysis, as well as nonlinear logit, probit, and Weibull methods, spelled out in Chapter 2, Section 7. We examine two cases, one for classification of credit card default using German data, and the other for banking intervention or closure, using data from Texas in the 1980s. Both of these data sets and the results we show are solely meant to be examples of neural network performance relative to more traditional econometric methods. There is no claim to give new insight into credit card risk assessment or early warning signals for a banking problem.

We see in both of the examples ...