Algoritmo de random forest aplicado a la detección de fraude en el sistema bancario ecuatoriano

Abstract

The purpose of this study is the construction of a statistical model, through which the probability that the client of a banking institution has to carry out fraud after granting some type of credit is estimated. For the elaboration of the model a modern methodology is contemplated based on measures of association, divergence measures, resampling and the machine learning technique known as Random Forest. In addition, the methodology used for the construction of the model will be implemented in the statistical software R, thus allowing to obtain the results automatically and saving a great amount of time.