Wind Speed Prediction Based on Multilevel Wavelet Decomposition Network
Jinxin Liu, Bin Qin, Leihuai Kuang, Xiaodong Xie · 2021 IEEE 5th Information Technology,Networking,Electronic and Automation Control Conference (ITNEC) · 2021
A Multi-resolution Recurrent Neural Network (MRNN) model based on multi-level wavelet decomposition network is proposed. The model incorporates wavelet decomposition algorithms into neural networks. Thus, an end-to-end ultra-short-term wind speed prediction is achieved. The MRNN model outperformed other comparison models in both one-step prediction and three-step prediction. And the MRNN model also has the optimal MAE evaluation index in the five-step prediction, which shows the advantages of MRNN in the prediction. Then this paper nested the ARIMA model on the basis of the MRNN model to further improve the prediction accuracy of the model. It can be considered as an in-depth exploration of time series prediction.