将小波多分辨率分析特点和支持向量机良好的泛化性能相结合,建立小波-支持向量机风速预测模型。先将原始风速序列经小波分解成概貌分量和细节分量,再对各分量分别应用支持向量机模型进行预测,最后将各分量的预测结果经小波重构得到原始风速序列的预测值。仿真表明该方法能够改善预测滞后现象以及减小突变点误差,从而提高模型的泛化性能和预测精度。 Combining the characteristics of wavelet multiresolution analysis with the good generalization performance of SVM, the wind speed prediction model of wavelet-support vector machine is established. Firstly, the original wind speed sequence is decomposed into the profile components and the detail components by wavelet decomposition, and each component is predicted by the support vector machine model. Finally, the prediction results of the components are reconstructed by wavelet to get the predicted value of the original wind speed sequence. Simulation results show that this method can improve the prediction lag and reduce the mutation point error, so as to improve the generalization performance and prediction accuracy of the model.