Commercial Mortgage Default: A Comparison of Logit with Radial Basis Function Networks

Abstract

The use of artificial neural networks in the modeling of foreclosure of commercial mortgages is explored by employing a large set of individual loan histories previously used in the literature of proportional hazard models, and specifically in Vandell et al (JAREUEA: 21(4), 451-480). Radial basis function networks are trained on the same input variables as those used in the logistic model. The networks are shown to be superior to the logistic benchmark in terms of discriminating between "good" and "bad" mortgages. The paper presents a sensitivity analysis on the average loan and offers suggestions for improving defaulting loan prediction.

Note: This is a description of the article and is not the actual abstract.

JEL Classification: R0, G21, G12