An LSTM Wind Power Prediction Model Based on PCA Dimensionality Reduction and Wavelet Decomposition

This paper introduces a wind power prediction method based on LSTM neural network. The method uses principal component analysis (PCA) to downscale climate data, and then uses wavelet decomposition to divide the processed climate data and historical power data into multiple frequency band signals. For each frequency band sequence, the LSTM model is used for fitting and prediction respectively. Finally, the overall wind power prediction results are obtained by integrating the predicted power signals from each frequency band. This method aims to improve the accuracy and reliability of wind power prediction.