摘要：The on-going European political and economic integration expands the scope of the target population of credit applicants from a single country to the European Union. Traditionally, accept/reject decisions in consumer credit are based on credit scoring models empirically derived for a specific country. The paper addresses the possibility of applying a single generic model to score the population of several European countries and the impact of segmentation (i.e. building individual models for different nations) on predictive accuracy. In contrast to previous studies, the logistic regression generic model built for three European countries shows the adequate performance, comparable to that of the national models.
Loading...
