Bibliographic citations
This is an automatically generated citacion. Modify it if you see fit
Diaz, K., (2024). Modelo predictivo de exportación de café utilizando Data Mining para los exportadores peruanos [Tesis, Universidad Privada del Norte]. https://hdl.handle.net/11537/38997
Diaz, K., Modelo predictivo de exportación de café utilizando Data Mining para los exportadores peruanos [Tesis]. PE: Universidad Privada del Norte; 2024. https://hdl.handle.net/11537/38997
@mastersthesis{renati/1691082,
title = "Modelo predictivo de exportación de café utilizando Data Mining para los exportadores peruanos",
author = "Diaz Preciado, Katherine Elizabeth",
publisher = "Universidad Privada del Norte",
year = "2024"
}
Title: Modelo predictivo de exportación de café utilizando Data Mining para los exportadores peruanos
Authors(s): Diaz Preciado, Katherine Elizabeth
Advisor(s): Castañeda Vargas, Pedro Segundo
Keywords: Exportaciones; Café; Minería de datos; Predicción; Competitiveness
OCDE field: https://purl.org/pe-repo/ocde/ford#2.02.04
Issue Date: 26-Sep-2024
Institution: Universidad Privada del Norte
Abstract: Las exportaciones de comodities son el pilar del crecimiento económico para el Perú, se ven afectadas por un entorno global dinámico y competitivo y los exportadores peruanos carecen de herramientas precisas para anticipar las fluctuaciones del mercado; por lo tanto, el propósito de la tesis es construir un modelo predictivo de exportación de café, utilizando técnicas de Data Mining. Se empleó una metodología cuantitativa de nivel descriptivo-explicativo, con un diseño no experimental y método inductivo; se hizo uso de los datos diarios de exportación de café publicados en la web de SUNAT del periodo 2013-2023, teniendo un total de 48,042 registros analizados, creando el dataset con los campos de: día, mes y fecha de embarque y peso neto; se utilizó la metodología CRISP-DM en conjunto con el modelo estadístico ARIMA con frecuencia mensual, donde se identificó los patrones de comportamiento que ha tenido la exportación de café peruano. Como resultados se evidenció el impacto que se tuvo durante la pandemia de COVID-19, es por lo que se realiza dos segmentaciones en el análisis y resultados obtenidos: antes y durante la pandemia. En cuanto a la validación, se aplicó el Test de raíz unitaria de Dickey-Fuller para validar que el comportamiento del peso neto de las exportaciones nacionales de café luego de la transformación cumpla con el requisito estacionariedad; mediante la función AC y ACP se encontró el número optimo de coeficientes del modelo AR y MA (2,1,3) estaba dentro de los intervalos de confianza, también se evaluó que el error presente comportamiento de ruido blanco mediante el test de Pormanteu, asimismo se evaluó que el comportamiento de ruido blanco en la función AC y PAC este dentro de los intervalos, encontrando que el modelo tuvo un porcentaje de significancia del 95%. En la validación del modelo predictivo se encontró un MAPE de 17.70%, mostrando que el modelo presenta una alta precisión en sus pronostico; el modelo predice el comportamiento futuro de 24 temporadas (2024-2025) que permitirá a los exportadores peruanos de café anticipar las tendencias del mercado internacional de café, mejorando así su capacidad para competir en un entorno muy dinámico, donde facilita la planificación estratégica y optimiza las operaciones relacionadas con la producción y preparación del café destinado al mercado exterior.
Commodity exports are the pillar of economic growth for Peru, they are affected by a dynamic and competitive global environment and Peruvian exporters lack accurate tools to anticipate market fluctuations; therefore, the purpose of the thesis is to build a predictive model of coffee exports, using Data Mining techniques. A quantitative methodology of descriptive-explanatory level was used, with a non-experimental design and inductive method; use was made of daily coffee export data published on the SUNAT website for the period 2013-2023, having a total of 48,042 records analyzed, creating the dataset with the fields of: day, month and date of shipment and net weight; the CRISP-DM methodology was used in conjunction with the ARIMA statistical model with monthly frequency, where the patterns of behavior that has had the export of Peruvian coffee were identified. The results showed the impact during the COVID-19 pandemic, which is why two segmentations were made in the analysis and results obtained: before and during the pandemic. As for the validation, the Dickey-Fuller unit root test was applied to validate that the behavior of the net weight of national coffee exports after the transformation complies with the stationarity requirement; The optimal number of coefficients of the AR and MA model (2,1,3) was found to be within the confidence intervals by means of the AC and ACP function; it was also evaluated that the error presents white noise behavior by means of the Pormanteu test, and that the behavior of white noise in the AC and PAC function is within the intervals, finding that the model had a percentage of significance of 95%. In the validation of the predictive model a MAPE of 17.70% was found, showing that the model presents a high precision in its forecasts; the model predicts the future behavior of 24 seasons (2024-2025) that will allow Peruvian coffee exporters to anticipate the trends of the international coffee market, thus improving their capacity to compete in a very dynamic environment, where it facilitates strategic planning and optimizes the operations related to the production and preparation of coffee destined for the foreign market.
Commodity exports are the pillar of economic growth for Peru, they are affected by a dynamic and competitive global environment and Peruvian exporters lack accurate tools to anticipate market fluctuations; therefore, the purpose of the thesis is to build a predictive model of coffee exports, using Data Mining techniques. A quantitative methodology of descriptive-explanatory level was used, with a non-experimental design and inductive method; use was made of daily coffee export data published on the SUNAT website for the period 2013-2023, having a total of 48,042 records analyzed, creating the dataset with the fields of: day, month and date of shipment and net weight; the CRISP-DM methodology was used in conjunction with the ARIMA statistical model with monthly frequency, where the patterns of behavior that has had the export of Peruvian coffee were identified. The results showed the impact during the COVID-19 pandemic, which is why two segmentations were made in the analysis and results obtained: before and during the pandemic. As for the validation, the Dickey-Fuller unit root test was applied to validate that the behavior of the net weight of national coffee exports after the transformation complies with the stationarity requirement; The optimal number of coefficients of the AR and MA model (2,1,3) was found to be within the confidence intervals by means of the AC and ACP function; it was also evaluated that the error presents white noise behavior by means of the Pormanteu test, and that the behavior of white noise in the AC and PAC function is within the intervals, finding that the model had a percentage of significance of 95%. In the validation of the predictive model a MAPE of 17.70% was found, showing that the model presents a high precision in its forecasts; the model predicts the future behavior of 24 seasons (2024-2025) that will allow Peruvian coffee exporters to anticipate the trends of the international coffee market, thus improving their capacity to compete in a very dynamic environment, where it facilitates strategic planning and optimizes the operations related to the production and preparation of coffee destined for the foreign market.
Link to repository: https://hdl.handle.net/11537/38997
Discipline: Maestría en Ingeniería de Sistemas con Mención en Gerencia de Sistemas de Información
Grade or title grantor: Universidad Privada del Norte. Escuela de Posgrado y Estudios Continuos
Grade or title: Maestro en Ingeniería de Sistemas con Mención en Gerencia de Sistemas de Información
Juror: Mendoza de los Santos, Alberto Carlos; Paredes Vargas, Ronal Santos; Lomparte Alvarado, Romulo Fernando
Register date: 19-Dec-2024
This item is licensed under a Creative Commons License