Abstract:
|
Financial institutions use risk models for measure both credit granting and customer behavior. The possibility of assigning a loan to a person could be assessed through regional economic variables that characterize where the customer comes from. In this work we propose to use a spatial analysis based on lattice models to identify factors that help to build a customer payment profile. To evaluate the performance of this model we compare it with a model which does not include regionalized spatial information. Finally, we illustrate the spatial statistical model using a real database, with credit information, from one important financial institution in Colombia.
|