Credit risk modeling for large datasets

Paulo H. Ferreira (UFBA)

Abstract: Statistical methods have been widely employed to assess the capabilities of credit scoring classification models in order to reduce the risk of wrong decisions when granting credit facilities to clients. Since massive amounts of financial data are widely available nowadays, we need to use appropriate statistical techniques, which, in general, deal with this problem by working with fewer subsets/samples of the whole database. Among these techniques, we highlight the segmentation procedure based on the client value, as well as the classification modeling with state-dependent sample selection. A discussion on these procedures, as well as applications to real datasets, will be provided in this presentation.