Developing interpretable data mining decision models for credit risk

Baesens, Bart (2006) Developing interpretable data mining decision models for credit risk. Workshop on Feature Selection in Data Mining, Oxford, UK. 28 Jun 2006.

Record type: Conference or Workshop Item (Paper)

Abstract

The recent introduction of the Basel II Capital accord has increased the need to develop data mining based decision models for credit risk. Since these models will play a pivotal role in the key business areas of a financial institution (e.g. credit granting, regulatory capital calculation, securitization, …), they need to be powerful but also interpretable. In this presentation, we will elaborate on how complex data mining techniques (e.g. neural networks and support vector machines) can be efficiently used to develop high performing and white-box decision models for financial risk management. We will hereby extensively discuss the importance of feature selection, rule extraction, two-staged models and knowledge representation.

This record has no associated files available for download.