WP 8 - Efficient and accurate risk prediction for credit operations

 

OBJECTIVES

  • To provide a solution to the credit risk prediction problem using hybrid Bayesian networks.
  • To develop a risk profiling methodology that can be used to improve marketing campaigns in a financial institution
  • To develop a prototype software system that will incorporate the specific features for credit risk prediction and profiling.
 

PLANNED WORK

In WP8 the general AMIDST framework developed in WP2-4 will be instantiated to the problem of credit risk assessment. This way, the methodological advances in variable selection, classification, MAP inference and learning algorithms for hybrid domains will be applied to a dataset provided by Cajamar and will be evaluated to check if they improve current credit risk prediction methods and help to design better marketing campaigns.

Also, a software system incorporating the specific features for the credit risk prediction use-case will be built.
 


This website uses cookies. Cookies are only used for handling login and identifying registrered users, not for visitor tracking.