Previsão da Inadimplência através da Regressão Logística

Abstract

With the progressive growth in volumes of lending to micro and small companies, financial institutions are looking increasingly for agility and assertiveness in granting credit. As the loans are granted, there is a possibility of loss, the ability to estimate the probability of occurrence of this loss makes the most reliable credit decision process. The objective of this work is to propose the use of a logistic regression model to estimate this probability of loss. The steps to build the model are discussed in detail, being approached from planning the choice of variables to the diagnostic model fit. At the end of work is presented a case study with the preparation of the logistic regression model to predict the probability of default customers, micro and small companies, a financial institution. The results of this particular model were analyzed and found to be effective to aid in the decision to grant credit for this customer profile.