Análise estatística multivariada: avaliando a (in)solvência de empresas comerciais utilizando regressão logística

Abstract

Nowadays, companies offering credit seek increasingly using statistical tools to combat delinquency in their transactions. The correct analysis of credit always presented himself as a key factor for institutions that provide. In this sense, the objective of this study is to generate a model of Credit Scoring (statistical model for the evaluation of credit) employing the techniques of factor analysis and logistic regression to assess the ability of classification of insolvency and solvency to commercial enterprises. For these techniques, we have the parameters regarding the degree of efficiency of the model generated and applied to commercial enterprises. It was also observed that the method of factor analysis helped the logistic regression, the accuracy of a hit for the classification. It can be concluded that after the use of factorial analysis the logistic regression obtained accuracy accuracy of 92% in the standings of commercial enterprises, which validates the model. It can be said that the model generated is appropriate for the classification of credit risk.