This paper explores the properties of using a generalized additive model with embedded variable selection for the prediction of bankruptcy. The main purpose is to explore an innovative way to close the gap between interpretation and prediction that has prevented widespread use of methods based...