Examinando por Materia "Series de tiempo"
Mostrando 1 - 19 de 19
Resultados por página
Opciones de ordenación
Ítem A retail demand forecasting system of product groups characterized by time series based on “ensemble machine learning” techniques with feature enginnering(Universidad EAFIT, 2022) Mejía Chitiva, Santiago; Aguilar Castro, José LisandroÍtem Aplicaciones prácticas de los métodos estadísticos para pronósticos en series de tiempo(Universidad EAFIT, 1993) Sánchez A., Javier; Universidad EAFITÍtem Aproximación del uso de redes neuronales en mantenimiento(Universidad EAFIT, 2013) Orozco Álvarez, David; Mora Gutiérrez, Luis AlbertoÍtem Cotton Price Long-Term Time Series Forecasting : A look at Transformers Suitability(Universidad EAFIT, 2024) Salazar Escobar, Carlos Enrique; Olarte, TomásRecent years have witnessed a surge of Transformer-based models for long-term time series forecasting (LTSF). These models boast impressive results in Natural Language Processing (NLP) and Computer Vision (CV), but their effectiveness in capturing the crucial temporal order inherent in time series data remains a question. This work investigates the suitability of Transformer-based models for long-term commodity price prediction, by replicating the work presented in "Are Transformers Effective for Time Series Forecasting?" by Zeng et al. (2022). We aim to evaluate their effectiveness compared to simpler baselines and analyze their limitations in capturing long-range dependencies. By delving deeper into these limitations, this research seeks to contribute to the development of more effective forecasting models for commodity price prediction.Ítem Deep Learning como alternativa en la predicción del precio de las acciones del mercado de valores colombiano(Universidad EAFIT, 2021) Uribe Ramírez, Sebastián; Almonacid Hurtado, Paula MaríaÍtem Demand forecasting in a manufacturing company, comparing traditional statistical and artificial intelligence models(Universidad EAFIT, 2022) Vásquez Jaramillo, Daniela; Castro Zuluaga, Carlos Alberto; Almonacid Hurtado, Paula MaríaÍtem Desarrollo de una metodología que permita determinar en forma previa la condición AR.I.MA. de una o múltiples series de tiempo, en un programa de base excel, para predicciones e inventarios en mantenimiento(Universidad EAFIT, 2012) Plaza Sibaja, Oscar Emilio; Mora Gutiérrez, Luis AlbertoLa metodología de series de tiempo, analiza las características de los datos del presente y del pasado, para proyectarlas hacia el futuro, donde se infiere que las causas que originan el comportamiento en el pasado y en el presente, son los mismos que condicionan el comportamiento futuro (Makridakis, y otros, 1978). La metodología de series temporales ofrece unos niveles de precisión entre lo predicho y la realidad cercanos e inferiores al 11%. Su metodología se basa en los principios de desarrollo del método científico: observación y análisis, hipótesis y verificación (Carrión, 1999). La hipótesis normal de los modelos proyectivos con múltiples variables, es que las variables no se relacionan entre sí, lo que se puede asumir como una limitación a este método futurístico, pero de todas maneras a pesar de esta condición, son útiles. La gran utilidad de los modelos proyectivos de series temporales es cuando se usan para estudios de una sola variable y cuando de alguna manera se desconoce las causas que los imputan, pues en ese caso donde se tenga claridad de cuáles son las variables que los afectan, más bien se estudia el futuro de estas causas, que el de la variable efecto primaria (Mora, 2006). El concepto de serie temporal se define como un conjunto de datos obtenidos del análisis y de las observaciones de una variable discreta durante un lapso secuencial de tiempo, es importante recordar que existen datos no temporales, son observaciones que se realizan de una forma no hilada en el tiempo. La serie de tiempo es un conjunto de datos de una variable, que se asocia a otro grupo de instantes definidos de tiempo; lo que implica el estudio de dos variables, donde una de ellas es el tiempo y la otra representa el fenómeno que se desea pronosticar (Bas, 1999). Los repuestos de mantenimiento presentan una demanda histórica baja, lo que traduce esto en series de tiempo con presencia de valores de cero, lo que dificulta el análisis de estos datos por los métodos determinísticos clásicos de la metodología de series temporales, requiriendo el uso de modelos genéricos no determinísticos como lo son los modelos AR.I.MA. (Díaz, 1991). La necesidad de disponer de predicciones lo más precisas posibles además del interés en conocer la dinámica de las distintas variables, origina que los métodos de análisis de series de temporales ocupen un lugar central en el estudio de disciplinas y fenómenos muy diversos (Peiró, y otros, 2000).Ítem Desarrollo de una metodología que permita determinar en forma previa la condición AR.I.MA. de una o múltiples series de tiempo, en un programa de base Excel, para predicciones e inventarios en mantenimiento(Universidad EAFIT, 2012) Plaza Sibaja, Oscar Emilio; Mora Gutiérrez, Luis AlbertoÍtem ¿Es posible pronosticar el precio por kilogramo en el mercado porcícola como una herramienta de gestión de riesgo?(Universidad EAFIT, 2024) Zapata Bustamante, Juan Camilo; Cruz Castañeda, VivianDue to the high volatility on the price per kilogram (COP/kg) in the pork market, the development of a predictive model as a risk management tool for producers is sought. The purpose of this tool is to provide a strategic guide to identify the best moment to sell a batch, organize the production accordingly and stablishing favorable conditions in selling contracts. This could provide an optimal risk management tool from the producer’s perspective in the market. To achieve this, the Box Jenkins methodology will be employed, using the ARIMA model as a base. The main objective its to anticipate possible fluctuations in the pig market, allowing producers to take informed decisions and in consequence, maximize de returns in the Colombian market operations.Ítem Evaluación financiera y recomendación gerencial para la reactivación del campo Libertad y la Estación de Tratamiento de Crudo Estrella (ETCE)(2019) Cubides Cardona, Juan Sebastián; Urrea Morales, Jair Leonardo; Waserman Álvarez, Jean PaulThis document discloses the recommendations given by the authors to the company, Oil Company S.A., as a result of the financial valuation of three different technical alternatives proposed for the restarting of the Libertad oil field and Estrella crude oil treatment station. The authors are based on financial concepts and tools, such as: 1. Valuation, which allows to build projections and financial planning, 2. Cash flow method, using the cost of weighted average capital that, as an advantage, contemplates free cash flow discounting the costs and expenses incurred in the operation, 3. Econometric models, that allows the estimation of high impact variables. This document will be part of the tools used by the management of the Libertad field, in order to make them able to support the final choice that generates the most value, and in this way focus all its resources on a short, medium and long term strategic plan that help them to achieve the objectives of the company.Ítem Metodología para estimación de series de tiempo en bonos : observación de volatilidad en bonos de deuda pública colombianos (TES)(Universidad EAFIT, 2019) Montoya López, Andrés; Ocampo Marín, Daniel Alberto; Mora Cuartas, Andrés Mauricio; Almonacid Hurtado, Paula MaríaThe application of time series methodologies in bond prices turns out to be a complex process due to its mathematical and financial characteristics. In particular, the coupon payment and the modified duration effect makes prices (clean and dirty) unable to be used correctly in time series applications. Academic community has proposed many different methodologies that could be applied to reduce the bias when estimating volatility on bond prices. It’s necessary to evaluate advantages and disadvantages on each of them, and analyze the characteristics that must have the time series. The investigation aims to propose a new methodology to estimate a time series that allow the modeling of Colombian public debt bond prices, observing its implementation in the volatility estimation.Ítem Modelamiento predictivo del número de visitantes en un centro comercial(Universidad EAFIT, 2022) Rua Jaramillo, Ramón David; Laniado Rodas, Henry; Almonacid Hurtado, Paula MaríaThe ability to make predictions about the number of customers or visitors in a shopping center is a very important input in the planning and efficient use of physical and human resources in this type of company. Also, it is important to understand what aspects influences their behavior. Based on historical data on the number of visitors, as well as external (environment) variables and online search trends, a forecasting model of the behavior of daily visits to the shopping center is suggested. The historical data correspond to the pedestrian and vehicular entries (cars and motorcycles) of the last 6 years in a shopping center located in the city of Medellín. This project begins with a literature review regarding forecasting models in different places such as museums, airports, natural parks, shopping centers and restaurants, among others, in order to explore methodologies in such cases and possible solution options. Through time series analysis and machine learning algorithms, the most representative variables and the best-fit model are selected to predict the number of visitors. This model is expected to be strengthened with estimation algorithms, improving performance over time and allowing it to be applied in other business or educational environments.Publicación Modelo AR-MIMO : mejora del pronóstico de múltiples horizontes en series de tiempo con optimizaciones heurísticas(Universidad EAFIT, 2025) Arias Zuluaga, Pablo Simón; Saldarriaga Aristizábal, Pablo AndrésÍtem Nueva Metodología Para Clasificar Datos de Series Temporales usando el Algoritmo Biclustering(Universidad EAFIT, 2013) Cogollo F. M.; Palacios, Alejandro; Universidad EAFIT. Escuela de Ciencias. Grupo de Investigación Modelado MatemáticoPublicación Predicción de ventas para una empresa de Hardware Business-to-Business(Universidad EAFIT, 2025) Sánchez Cárdenas, Hernán Felipe; Almonacid Hurtado, Paula MaríaÍtem Predicción del cargue de rutas de distribución mediante aprendizaje de máquina(Universidad EAFIT, 2023) Ramírez Aguilar, Santiago; Téllez Falla, Diego Fernando; Marentes Cubillos, Luis AndrésÍtem Predicción del precio de transacción sobre el tipo de cambio XAU-USD (oro) para el mercado de contado del commodittie a corto plazo(Universidad EAFIT, 2023) Cardona Restrepo, Jorge Esteban; Castilla Rueda, Rafael Andrés; Almonacid Hurtado, Paula MaríaÍtem Uso de kernels en series tiempo para la detección de prácticas manipulativas en mercados financieros(Universidad EAFIT, 2023) Herrera Ochoa, José Daniel; Quintero Montoya, Olga LucíaIntuitively, one might think that any deviation in trading data could be easily detected due to the statistical basis on which finance sciences are based. However, the markets in which financial assets are traded operate under the principle of supply and demand, as well as the principle of opportunity. Elements that make them very susceptible to price manipulation. For this reason, it is increasingly relevant to consider techniques that allow the identification of elements in financial time series that can deliver information that show whether a stock has been subject of manipulative practices or not. The use of kernels for signals decomposition and filtering in financial time series is then proposed. By using this technique elements of the time series such as power and frequency can be obtained, which can later facilitate the characterization of a stock that has been subject of fraudulent or manipulative trading. Then considering diverse machine learning techniques, achieve a timelier detection based on said characterization, particularly in dynamic and constantly evolving trading environments. For this purpose, the performance of the kernels will be contrasted against traditional techniques, choosing the most appropriate ones. In the same way, various machine learning techniques will be evaluated and the one that best learns and represents the patterns or artifacts in fraudulent operations will be chosen. Trying in this way to raise trading standards in financial markets, as well as delving into the applications that the decomposition and filtering of signals with kernels can have, not only as a data visualization tool, but also as inputs. for machine learning techniques.Ítem Volatilidad de precios y fuentes de energía renovable no convencionales : el caso colombiano(Universidad EAFIT, 2020) Cardona Vásquez, David; Arango Manrique, Adriana; García Rendón, Jhon Jairo