期刊名称：Revista Innovar Journal Revista de Ciencias Administrativas y Sociales
印刷版ISSN：2248-6968
出版年度：2010
卷号：20
期号：36
页码：7-14
语种：Spanish
出版社：Universidad Nacional de Colombia - Sede Bogotá
摘要：Contracts in unregulated electricity markets are a tool to protect the agents from volatility; in this conlext, price predictions are a key input to enable customers to make strategic and operational decisions. In this article, average prices are predicted fur contrart, signed in the Colombian electricity market, using a neuronal network whith dinamic architecture known as DAN2. The model developed is able to capture the intrinsic dinamic of series of prices and forecast the price for the following month with greater precision than the classic ARIMA methodology for forecasting horizons of 12 and 24 months.
关键词：Electricity prices;artificial neuronal networks;prediction;time series;precios de la electricidad;redes neuronales artificiales;predicción;series de;Prix de l´électricité;réseau neuronal artificiel;prédiction;series de témps;preços da eletricidade;redes neuronais artificiais;prediçao;series de tempo
