Short-term wind power prediction based on kpca-kmpmr

Wind power prediction is significant for power system dispatching and safe-stable operation. This paper proposes a new approach for wind power prediction. It is derived by integrating the kernel principal component analysis (KPCA) method with a new probability learning method. kernel minmax probability machine regression (KMPMR). In the proposed model. KPCA is used to extract features of the inputs and obtain kernel principal components. Then. KMPMR is employed to predict the short-term wind power by using the real dataset from wind farms of Alberta. Canada. The results show that. under the same conditions. the proposed method provides a better prediction performance than KMPMR and PCA-KKMPMR methods.