Reducing Electricity Price Forecasting Error Using Seasonality and Higher Order Crossing Information

Commodity prices are volatile and their volatility changes over time. Electricity prices are even more volatile in general because customers cannot easily switch to other energy source when the electricity prices fluctuate. Electricity prices also have strong seasonal behavior because the demand of electricity depends on the season. To address these characteristics, a time series forecasting model with seasonality and higher order crossing information for electricity prices is proposed. The model measures the similarity between two time series by using their higher order crossing information. The model can predict not only day-ahead prices (short term forecasting), but also a series of prices for one week ahead or even longer simultaneously (medium term price-curve forecasting). The model is tested on both high and low volatile data sets to evaluate its robustness and generality. It is also shown that the model can produce more accurate prediction than some well-known models.