Bibliographic citations
This is an automatically generated citacion. Modify it if you see fit
Bernal, J., (2020). Estimación de las ventas procesadas de la financiera PERUNET aplicando redes neuronales y SARIMA 2019 [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/21811
Bernal, J., Estimación de las ventas procesadas de la financiera PERUNET aplicando redes neuronales y SARIMA 2019 [Tesis]. PE: Universidad Nacional de Ingeniería; 2020. http://hdl.handle.net/20.500.14076/21811
@misc{renati/711479,
title = "Estimación de las ventas procesadas de la financiera PERUNET aplicando redes neuronales y SARIMA 2019",
author = "Bernal Fernández, Julio César",
publisher = "Universidad Nacional de Ingeniería",
year = "2020"
}
Title: Estimación de las ventas procesadas de la financiera PERUNET aplicando redes neuronales y SARIMA 2019
Authors(s): Bernal Fernández, Julio César
Advisor(s): Chíncaro Del Coral, Omar Antonio
Keywords: Predicción; Sarima; Redes neuronales artificiales; Series de tiempo
OCDE field: http://purl.org/pe-repo/ocde/ford#1.01.03
Issue Date: 2020
Institution: Universidad Nacional de Ingeniería
Abstract: En la actualidad muchas empresas u organizaciones para definir sus estrategias necesitan pronosticar el futuro de sus actividades en función del comportamiento del pasado con lo cual definen sus presupuestos. A partir de ello buscan minimizar los errores en la predicción haciendo uso de las técnicas estadísticas, para lo cual los modelos SARIMA y REDES NEURONALES ARTIFICIALES han sido utilizadas exitosamente en muchos tipos de problemas de series de tiempo, debido a que son capaces de modelar y predecir series estadísticas de tiempo lineal y no lineal con alto grado de precisión. El objetivo de la investigación es proporcionar una metodología que mejore los errores de pronósticos en la predicción de ventas que procesa la financiera PeruNet en donde se aceptan tarjeta de crédito y/o débito ya sea Visa, MasterCard u otra marca de tarjeta. La principal aportación es demostrar que el modelo de REDES NEURONALES es capaz de obtener buenas aproximaciones, tanto en el ajuste como en el pronóstico, además de observar que, en este caso en particular, la aproximación resulta mejor que la generada por medio de la metodología SARIMA e inclusive mejor que el pronóstico elaborado por la metodología empírica que realizan los especialistas de la financiera PeruNet.
Currently, many companies or organizations to define their strategies need to forecast the future of their activities based on the behavior of the past with what they define their budgets. From this, they seek to minimize the errors in the prediction by making use of statistical techniques, for which the SARIMA and ARTIFICIAL NEURONAL NETWORKS models have been used successfully in many types of time series problems, because they are capable of modeling and predict statistical series of linear and non-linear time with a high degree of precision. The objective of the research is to provide a methodology that improves forecast errors in the sales prediction processed by a financial Z where credit and / or debit cards are accepted, be it Visa, MasterCard or another brand of card. The main contribution is to demonstrate that the NEURONAL NETWORKS model is capable of obtaining good approximations, both in the adjustment and in the forecast, in addition to observing that, in this particular case, the resulting approximation is better than that generated by means of the methodology SARIMA and even better than the forecast prepared by the expert judgment of the employees of the financial Z.
Currently, many companies or organizations to define their strategies need to forecast the future of their activities based on the behavior of the past with what they define their budgets. From this, they seek to minimize the errors in the prediction by making use of statistical techniques, for which the SARIMA and ARTIFICIAL NEURONAL NETWORKS models have been used successfully in many types of time series problems, because they are capable of modeling and predict statistical series of linear and non-linear time with a high degree of precision. The objective of the research is to provide a methodology that improves forecast errors in the sales prediction processed by a financial Z where credit and / or debit cards are accepted, be it Visa, MasterCard or another brand of card. The main contribution is to demonstrate that the NEURONAL NETWORKS model is capable of obtaining good approximations, both in the adjustment and in the forecast, in addition to observing that, in this particular case, the resulting approximation is better than that generated by means of the methodology SARIMA and even better than the forecast prepared by the expert judgment of the employees of the financial Z.
Link to repository: http://hdl.handle.net/20.500.14076/21811
Discipline: Ingeniería Estadística
Grade or title grantor: Universidad Nacional de Ingeniería. Facultad de Ingeniería Económica, Estadística y Ciencias Sociales
Grade or title: Ingeniero Estadístico
Juror: Álvarez Rojas, Cirilo; Pinedo Sánchez, Amélida; Vásquez Rodríguez, Rafael
Register date: 11-Apr-2022
This item is licensed under a Creative Commons License