Wind power prediction method based on EMD-KPCA-BiLSTM-ATT model

The invention provides a wind power prediction method based on an EMD-KPCA-BiLSTM-ATT model, and the method comprises the steps: firstly, carrying out the decomposition of experimental sample data through EMD, and obtaining a series of IMF components and residual components; secondly, calculating the contribution rate of each IMF component by using a KPCA algorithm, carrying out dimension reduction processing, and forming a new data set by feature data after dimension reduction; performing normalization processing on data in the new data set, and then dividing the data into a training set and a test set; then, learning training is conducted on the BiLSTM-ATT combined prediction model through training set data, prediction results are compared, hyper-parameters reaching the target accuracy rate are determined, and then an optimal prediction model is obtained; and finally, testing the optimal prediction model by using the test set data to obtain the wind power to be predicted, and evaluating the prediction effec