Thesis PROPUESTA DE MODELOS ADAPTATIVOS HÍBRIDOS Y NO-HÍBRIDOS PARA EL PRONÓSTICO DE VOLATILIDAD DEL PRECIO MENSUAL DEL COBRE MEDIANTE ALGORITMOS GENÉTICOS
Loading...
Date
2018
Authors
GARCÍA BOZZO, DIEGO IGNACIO
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Este trabajo busca realizar pronósticos de volatilidad para el mercado del cobre en frecuencia mensual, lo cual es de interés práctico para diferentes participantes tales como productores, consumidores, gobiernos e inversionistas.Teniendo en cuenta datos históricos desde 1990 y los principales determinantes de este mercado en específico, se propone un conjunto de modelos de series de tiempo, modelos no-paramétricos, y especificaciones híbridas de ambos. La capacidad de adaptación de estos modelos tanto en variables exógenas, parámetros de configuración y tamaño de ventana, simultáneamente, es entregada mediante un algoritmo genético con el propósito de alcanzar los mejores pronósticos posibles.El desempeño fuera de muestra examinado se basa en el Heteroskedasticity-adjusted Mean Squared Error (HMSE), y se evalúa la superioridad de modelos usando el Model Confidence Set (MCS). Los resultados muestran que realizar pronósticos de una forma adaptativa es crucial para obtener desempeños robustos y mejorados. La especificación del tipo Adaptive-GARCH-FIS demuestra poseer el mayor poder de pronóstico.
This thesis studies volatility forecasting in monthly terms for the copper market, which is of practical interest for different participants such as producers, consumers, governments, and investors.Using data since 1990 and the main drivers of this specific market, we propose a set of time series models, non-parametric models, and hybrid specifications of both. The adaptability characteristic of these models in exogenous variables, their configuration parameters and window size, simultaneously, are provided by a genetic algorithm in pursuit of achieving the best possible forecasts.We examine out-of-sample performance based on Heteroskedasticity-adjusted Mean Squared Error (HMSE), and we test model superiority using the Model Confidence Set (MCS). The results show that making forecasts in an adaptive manner is crucial to obtaining robust and improved performances. The Adaptive-GARCH-FIS specification yields the best forecasting power.
This thesis studies volatility forecasting in monthly terms for the copper market, which is of practical interest for different participants such as producers, consumers, governments, and investors.Using data since 1990 and the main drivers of this specific market, we propose a set of time series models, non-parametric models, and hybrid specifications of both. The adaptability characteristic of these models in exogenous variables, their configuration parameters and window size, simultaneously, are provided by a genetic algorithm in pursuit of achieving the best possible forecasts.We examine out-of-sample performance based on Heteroskedasticity-adjusted Mean Squared Error (HMSE), and we test model superiority using the Model Confidence Set (MCS). The results show that making forecasts in an adaptive manner is crucial to obtaining robust and improved performances. The Adaptive-GARCH-FIS specification yields the best forecasting power.
Description
Catalogado desde la version PDF de la tesis.
Keywords
ALGORITMOS GENETICOS , MODELOS ADAPTATIVOS HIBRIDOS , PRECIO DEL COBRE
