Bibliographic citations
Miñano, M., Rodríguez, J. (2017). Desarrollo de una aplicación informática basada en un modelo de mahine learning para mejorar la evaluación de préstamos crediticios [Tesis, Universidad Privada del Norte]. https://hdl.handle.net/11537/12294
Miñano, M., Rodríguez, J. Desarrollo de una aplicación informática basada en un modelo de mahine learning para mejorar la evaluación de préstamos crediticios [Tesis]. PE: Universidad Privada del Norte; 2017. https://hdl.handle.net/11537/12294
@misc{sunedu/3027974,
title = "Desarrollo de una aplicación informática basada en un modelo de mahine learning para mejorar la evaluación de préstamos crediticios",
author = "Rodríguez Castillo, Jorge Junior",
publisher = "Universidad Privada del Norte",
year = "2017"
}
ABSTRACT The present research work is focused on the study of a machine learning model to develop an informatic application, which allow to improve the evaluation of credit loans by providing a better analysis of the profitability and credit risk; which is used by the financial company that for reason of privacy of its data we will call Financiera Nuestro Crédito. The problema lies in determining how an informatic application based on a machine learning model contributes to improving the evaluation of credit loans. To solve this problema was developed an informatic application based on a logistic regression model which allows to perform the evaluation and prediction of credit loans through a simple interface, where the user put some principal features such as; el amount requested, interest rate, credit terms, marital status and the age of the applicant. The logistic regression model is divided in four main algorithms, the first one is a function for scaling data, the second one is an algorithm for prediction that is called sigmoid function, the third one is a function to reduce the cost of the model and the last one is the algorithm of optimization called the gradient descent. This informatic application under the use of the logistic regression model will achieve to increase the percentage of earned money, decrease the amount of lost money and decrease of the time average for the approval of credit loans. The results of the reasearch work indicate that with the development of this informatic application based on a logistic regression model, the percentage of earned money is increased, the amount of lost money is reduced by the evaluation of the credit loans and the average time spent to approve a credit loan is reduced, too. Besides that, the effectiveness of this informatic application give an aceptable percentage of prediction for the financial company. For these reason, it is concluded that this informatic application is very useful to Financiera Nuestro Crédito.
This item is licensed under a Creative Commons License