Price forecasting by ICA-SVM in the competitive electricity market

Price forecaster is one of the important aspects in the researches of the electricity price. But because of the special properties of electricity, the price of electricity is far more volatile than that of other relatively volatile commodities. Due to such significant volatility, it is difficult to make an accurate forecast for the spot market of electricity. To solve above problem, a hybrid algorithm for price forecasting is proposed. The procedure comprises two main steps. The first step is feature extract of price feature by independent component analysis that has special effect on extracting the latent source feature. The second step is regression modeling by support vector machine training with the refined features that can realize higher-dimensional nonlinear regression with good generalization ability. In this way, the two algorithms have combined, whose advantages have been made a full use. Case studies are applied to test the proposed model.