Performance of credit risk prediction models via proper loss functions
The performance of predictions models can be assessed using a variety of methods and metrics. Several new measures have recently been proposed that can be seen as refinements of discrimination measures, including variants of the AUC (Area Under the ROC curve), such as the H index. It is widely recognized that AUC suffers from lack of coherency especially when ROC curves cross. On the other hand, the H index requires subjective choices. In our opinion the problem of model comparison should be more adequately handled using a different approach. The main contribution of this paper is to evaluate the performance of prediction models using proper loss function. In order to compare how our approach works with respect to classical measures employed in model comparison, we propose a simulation studies, as well as a real application on credit risk data.
|Date of creation:||Jan 2014|
|Date of revision:|
|Contact details of provider:|| Postal: |
Web page: http://epmq.unipv.eu/site/home.html
More information through EDIRC
References listed on IDEAS
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Crook, Jonathan N. & Edelman, David B. & Thomas, Lyn C., 2007. "Recent developments in consumer credit risk assessment," European Journal of Operational Research, Elsevier, vol. 183(3), pages 1447-1465, December.
When requesting a correction, please mention this item's handle: RePEc:pav:demwpp:demwp0064. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Alice Albonico)
If references are entirely missing, you can add them using this form.