On Credit Scoring Estimation

Credit scoring methods became a standard tool of
banks and other financial institutions, direct
marketing retailers and advertising companies to
estimate whether an applicant for credit/goods will
pay back his liabilities. In this book we give a
short overview of credit scoring and its
quantitative methods. We investigate the usage of
some of these methods and their performance on a
real data set taken from a French bank. Our results
indicate that the methods used, namely the logistic
regression, multi-layer perceptron (MLP) and radial
basis function (RBF) neural networks give very
similar results, however, the traditional logit
model seems to outperform the other techniques. We
also describe RBF architecture and a simple RBF
program that we implemented in the statistical
computing environment XploRe.

Autorentext

RNDr. Karel Komorád, M.Sc. studied mathematics at Charles University in Prague and statistics at Humboldt-Universität zu Berlin and Freie Universität Berlin. He worked for several years as a scoring analyst at KB Bank, Prague. Currently he works as a consultant at everis Spain, S.L.

Klappentext

Credit scoring methods became a standard tool of banks and other financial institutions, direct marketing retailers and advertising companies to estimate whether an applicant for credit/goods will pay back his liabilities. In this book we give a short overview of credit scoring and its quantitative methods. We investigate the usage of some of these methods and their performance on a real data set taken from a French bank. Our results indicate that the methods used, namely the logistic regression, multi-layer perceptron (MLP) and radial basis function (RBF) neural networks give very similar results, however, the traditional logit model seems to outperform the other techniques. We also describe RBF architecture and a simple RBF program that we implemented in the statistical computing environment XploRe.