Exponential Smoothing models for energy forecasting

In this paper we address the issue of modelling and forecasting spot electricity* prices and energy l,oad demand. In particular, we model hourly time series for the Italian 'GM3, (G.rtore Mercato Elettrico) and Nordic Nord Pool markets. Exponential smoothing Holt-Winters methods are appropriate in this context because they are highly adaptable and robust tools to forecast different horizons. Moreover, since the main characteristics of those time series are the strong seasonal patterns they display, we take into account for daily and weekly cycles. Mean Absolute Percentage Error is evaluated to provide a forecasting performance measure for the proposed modelExIn this paper we address the issue of modelling and forecasting spot electricity* prices and energy l,oad demand. In particular, we model hourly time series for the Italian 'GM3, (G.rtore Mercato Elettrico) and Nordic Nord Pool markets. Exponential smoothing Holt-Winters methods are appropriate in this context because they are highly adaptable and robust tools to forecast different horizons. Moreover, since the main characteristics of those time series are the strong seasonal patterns they display, we take into account for daily and weekly cycles. Mean Absolute Percentage Error is evaluated to provide a forecasting performance measure for the proposed mode