Diseño de un modelo matemático para la calificación de clientes morosos en una entidad comercial mediante las metodologías de árboles de decisión, análisis discriminante y regresión logística

In the present work, a mathematical model is designed for the qualification of delinquent clients in order to optimize the management of collections in a commercial enterprise. To do this we developed two parametric models, Discriminant Analysis and Logistic Regression and a non-parametric model, th...

Full description

Bibliographic Details
Author: Ruiz López, Hugo Renán
Format: article
Status:Published version
Publication Date:2017
Country:Ecuador
Institution:Universidad Internacional del Ecuador
Repository:Repositorio Universidad Internacional del Ecuador
OAI Identifier:oai:repositorio.uide.edu.ec:37000/3647
Online Access:https://doi.org/10.33890/innova.v2.n7.2017.334
https://repositorio.uide.edu.ec/handle/37000/3647
Access Level:Open access
Keyword:riesgo de crédito; árboles de decisión; análisis discriminante; regresión logística; morosidad
credit risk; decision trees; discriminant analysis; logistic regression; delinquency
Description
Summary:In the present work, a mathematical model is designed for the qualification of delinquent clients in order to optimize the management of collections in a commercial enterprise. To do this we developed two parametric models, Discriminant Analysis and Logistic Regression and a non-parametric model, the Decision Trees. It determines the stability of each model, its power of discrimination and finally establish which of the methodologies allows better optimization of collection management.