One hour ahead prediction of wind speed based on data mining

Wind speed forecasting is very important to the utilization of wind energy in wind farm. In order to improve the forecast precision, a forecasting method based on empirical mode decomposition (EMD) and wavelet decomposition combine with least square support vector machine (LSSVM) is proposed in this paper. The wind speed time series was decomposed into several intrinsic mode functions (IMF) and the trend term. In order to reduce the nature of non-stationary, the high frequency band was decomposed and reconstructed by wavelet transform (WT). The different LSSVM models to forecast each IMF and trend term were built up. These forecasting results of each IMF and trend term were combined to obtain the final forecasting results. The simulation experiment shows the MAPE is 4.53% about wind speed forecasting and the prediction accuracy is improved considerably.