Electricity price forecasting: A hybrid wavelet transform and evolutionary-ANN approach

In a restructured power market, the forecasting of price of electricity has drawn attention of researchers for an accurate forecasting of the electricity price. Electricity price forecast provides important information to the electricity market managers and participants. However, electricity price is a complex signal due to its non-linear, non-stationary, and time variant behavior. In spite of performed research in this area, more accurate and robust price forecast methods are still required. In this article a novel technique has been proposed to forecast the electricity prices using wavelet transform and a Feed-Forward Neural Network trained by a Meta heuristic algorithm i.e. Invasive Weed Optimization technique (IWO). The wavelet transform has been used to decompose ill-behaved price series in a set of better constitutive series. Here we have used the data of electricity market of Australia in year 2005 and the reported results have been compared with the ANN, trained by back propagation algorithm.