The Application Of Wavelet Analysis And Artificial Neural Network Price Forecasting

The deregulation of power industry is now the world's trend. How to price the special commodity-electricity in the power market is essential for the smooth and fair market operation. So to use the historic data in predicting the future electricity market price is a very meaningful work.There are similar researches on the stock market and foreign currency market. But the electric power market has its own special characteristics. Wavelet transform is a new mathematical branch developed since I 980s'. Since it has the capacity to distinguish different data factors, wavelet analysis has been used widely in the serial data analysis. The main strength of wavelet analysis is that it can divide data-changing frequency infinitely and it is then possible to use according prediction method in the different frequency band. Artificial Neuro Network (ANN) uses specimen data to optimize network parameters and then simulates any possible mapping function. But its main shortcoming is that the building of ANN doesn't have any regular method.This article tries to combine the two mean's advantages and improve the prediction performance. The Chapter 1 describes the electricity pricing method in the power market; Chapter 2 describes the wavelet analysis' characteristics; Chapter 3 introduces ANN's applications and drawbacks; Chapter 4 uses practical market data to evaluate the performance.