ANÁLISE DISCRIMINANTE MÚLTIPLA DE TRÊS GRUPOS NA PREVISÃO DE INADIMPLÊNCIA DE PESSOA FÍSICA – Aplica

Código do trabalho: 
T_15_337
Resumo: 
This article aims to analyze, through a statistical model, a set of customers’ data, setting them into three groups: on-time payers, intermediaries and late payers. In order to reach this goal, the concepts of credit and consortium as well as the characteristics of statistical equations that are used by multiple discriminant analysis were specified. It was then demonstrated the application of calculations using data of a group of customers from a consortium company, institution which sells automotive consortium quotas. Based on such data, equations that incorporate multiple discriminant analysis calculation were done in order to separate the set of clients, watching how close this division is real, its degree of error and the feasibility of its application for future customer data.