Implementación de un modelo score para el otorgamiento en una cartera de microcrédito

Abstract

Financial institutions seek every day to reduce operating costs related to the granting of credit, for this purpose a score model is developed to achieve this objective. This work uses the methodologies of logistic regression and discriminant analysis, taking as a reference a database made up of 469,996 clients belonging to the microcredit modality, of which there are qualitative and quantitative variables. The objective is to find the best model that allows to have the best estimate and that shows benefits when having a score model. It is concluded that logistic regression is the model that best differentiates and predicts good customers from bad ones, managing to reduce the level of expected losses.