"Development and validation of forecasting models based on logistic regression"
Download the SAS macros and important background information
R. Muche, C. Ring, C. Ziegler:
Development and validation of prediction models based on logistic regression
An evaluation strategy and its implementation with SAS macros
Illustrated using the example of modelling the risk of early retirement after an inpatient rehabilitation measure
Shaker-Verlag, Aachen (2005)
On this page you can download the SAS macros for developing and validating prediction models based on logistic regression, which are described in the monograph above:
Depending on which SAS version you are using, you can obtain the macros for version 8, version 9.1 or version 9.4.
In addition, three theses can be downloaded which describe the special background of the work and the macros:
- Diploma thesis Ziegler: Programming the macros
- Diploma thesis Assfalg: Bootstrap validation
- Diploma thesis Habel: Shrinkage validation
With the change to SAS versions 9.1 and 9.4, some changes had to be made to the macros. You can find a record of these here:
- Ruthenberg report (pdf): Testing and changing the macro package for use in SAS version 9.1
- Report Müller (pdf): Testing and modifying the macro package for use in SAS version 9.4
The macros were extensively validated. The corresponding tests are also available here as a validation report (incl. documentation and SAS programmes).
Unfortunately, the arrangements with the LVA-Baden-Württemberg are such that the data for reproducing the analyses in the book cannot be offered here. For didactic reasons, we would have liked to make them openly accessible, but the data protection regulations in the area of pension insurance expressly do not permit this.
Prof Dr Rainer Muche
(for the authors)