Unstable regions in the scorecards’ input space
In: Symposium for Young Researchers 2007: Proceedings
Data mining algorithms become more and more popular to satisfy the Basle II requirements, like to predict the probability of default. Not all of these models can be understood easily from economical point of view, which involve the importance of stress tests. In this paper we try to map a retail credit scorecard’s input space to find regions where predictions can lead to significant differing results. Different definitions for similarity and prediction difference are examined to reach an economically and statistically simultaneously interpretable abstraction.
|This chapter was published in: Anna Francsovics (ed.) Symposium for Young Researchers 2007: Proceedings, , pages 181-186, 2007.|
|This item is provided by Óbuda University, Keleti Faculty of Business and Management in its series Proceedings Papers of Business Sciences: Symposium for Young Researchers (FIKUSZ) 2007 with number 181-186.|
|Contact details of provider:|| Postal: |
Web page: http://www.kgk.uni-obuda.hu
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:pkk:sfyr07:181-186. 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: (Alexandra Vécsey)
If references are entirely missing, you can add them using this form.