In credit business, banks are interested in information whether prospective consumers will pay back their credit or not. The aim of credit-scoring is to model or predict the probability that a consumer with certain covariates is to be considered as a potential risk. The dataset consists of 1000 consumer credits from a german bank. For each consumer the binary response variable "creditability" is available. In addition, 20 covariates that are assumed to influence creditability were recorded.

Because of the large number of categorical covariates, we moved the list of covariates and their encoding to a separate page. A summary of the relative frequencies is also given there.

The dataset Determining the solidness of borrowers via creditscoring can be downloaded in the following versions:

Uncompressed ASCII-File

kredit.asc

62 KB

• Fahrmeir, L. / Tutz, G. (1994): Multivariate Statistical Modelling Based on Generalized Linear Models. Springer, New York.
• Fahrmeir, L. / Hamerle, A. / Tutz, G. (1984, 1st ed.): Multivariate statistische Verfahren. de Gruyter, Berlin.
• Fahrmeir, L. / Hamerle, A. / Tutz, G. (1996, 2nd ed.): Multivariate statistische Verfahren. de Gruyter, Berlin.
• Fahrmeir, L. / Hamerle, A. (1981): Kategoriale Regression in der betrieblichen Planung. Zeitschrift für Operations Research, 44/B, 63-78.
• Feilmeier, M. / Fergel, I. / Segerer, G. (1981): Lineare Diskriminanz- und Clusteranalyseverfahren bei Kreditscoring-Systemen. Zeitschrift für Operations Research, 25/B, 25-28.