Classificação de clientes inadimplentes via regressão logística

Abstract

Due to the necessity to understand the defaulting customers profile and create more efficient collection strategies, the objective of this study conclusion is to classify default by a private card company's customers according to the customer's propensity to negotiate debt. The data were extracted from a collection center's database located in Belo Horizonte, which provides services to a national large bank institution. A logistic regression model was adjusted and a ranking of clients was created according to their propensity to negotiate debt.