2 869 855 libros electrónicos en 110 idiomas
¿No le conviene? No hay problema. Puedes devolver los artículos hasta 30 días
No se equivocará con un vale de regalo. El destinatario puede elegir cualquier producto de nuestra oferta.
Hasta 30 días para devoluciones
In today s business environment the risk management of consumer lending has become critical to protect the interest of both lenders and consumers. Following the presentation of the main issue in the consumer credit market and the introduction of credit scorecards, I analyze a Hungarian dataset of loans for personal needs. The analysis evaluates comparatively the prediction accuracy of several methods used in consumer credit scoring. I have documented the superior performance of logit and neural estimations. I also have found supporting evidence for the use of WOE as inputs in estimating predicted default probabilities. Following Siddiqui (2006), I have used a linear transformation to compute a traditional statistical scorecard. Developing easy to interpret statistical scorecards is essential when the lender has a legal obligation to explain its refusal of an application.