Examinando por Autor "Rodriguez Avellaneda, Angelica Lucia"
Mostrando 1 - 1 de 1
Resultados por página
Opciones de ordenación
- ÍtemModelo scoring para recuperar cartera de microcrédito(Fundación Universitaria Los Libertadores. Sede Bogotá., ) Rodriguez Avellaneda, Angelica Lucia; Santana Contreras, Juan CamiloIn the present work the construction of a Scoring model is exposed to determine the probability of payment of the client when it is in default, focused on the sector micro credit, looking for a tool with which the portfolio of in an efficient way, reducing costs and increasing the level of recovery. Finally If this objective is met, there is a 6-month base (April-September 2018), with which 4 different models are generated, which are: logistic regression, regression probit, decision trees and neural networks. The percentage of error in each one of them, resulting in the lower percentage generated by decision trees with 11.3%, said model reflects an adequate adjustment, showing a Successful predictability level. As for the remaining models, none have a percentage greater than 12.5%. Therefore, through logistic regression, it is determined that the guarantee, gender and default at the time of evaluation are the variables with the highest impact on non-payment of customers.