Predicción de incumplimiento de pago de clientes de tarjetas de crédito, con aplicación del algoritmo del k-vecino más cercano y Clas-FriedmanAligned-ST

Companies that give credit cards to clients face some problems such as non-payment, which is why companies need to control such debts, so as to minimize the risk of recovery of the investment, as a result of debtor clients. In this article, the lazy learning algorithm KNN with the method of statistical evaluation Clas- FriedmanAligned-ST was used, to help us to predict the degree of nonpayment of debts, in order to optimize and improve the prediction performed by data mining algorithms. The database used for this work contains 30000 records, each defined by 25 attributes, of which a significant sample of 5439 instances was taken, with 24 fields. A data processing model is developed, the results are discussed; And concludes with the benefits of evolutionary computing application.