The non-payment (breach of contract) is one of the major, if not the major, problem faced by administrators (companies, agencies) of credit. In studies of such problems it was created the risk concept, that is essentially the probability of not receiving the credits from the administrators. Some authors, Caouette et al. (2000) and Silva (1988) , refer the multivariate analysis as a very powerful tool in the risk administration of conceding the personal credit. This paper show the build and the evaluation of pattern recognition and classified rules based on the Discriminant Linear Function and the Logistic Regression, to classify the clients of credit card in one of two groups. The efficiency of the procedures was evaluated by the Lachenbruch Method, Lachenbruch (1975).
Pattern recognition; credit analysis; discriminant linear function; logistic regression; Lachenbruch's method
