Maestría en Administración Financiera · 2018
Estimación de un modelo del precio de la energía eléctrica en Colombia con detección de puntos de volatilidad, utilizando la transformada Wavelet y series de tiempo
In this document we propose a technique to estimate a model of the price of electric power in Colombia, using the filtering properties of the Discrete Wavelet Transform (DWT) and the traditional time series modeling techniques ARIMA, GARCH; in addition, the detection of points of change in the variance of the price series through the detection method of multiple changes in a sequence of dependent variables; the series of spot prices is decomposed in a series of approximation and several details, then each sub-series separately is modeled with the technique that best fits the data -- The final forecast is the sum of the reconstructed forecasts obtained from each sub-series -- The most important conclusion of the proposed model is to allow greater precision in the forecast and, in turn, detect points of volatility change due to exogenous variables in the series of spot prices of the Colombian electricity market