Spatio-Temporal Wind Speed Prediction Based on Variational Mode Decomposition

Improving short-term wind speed prediction accuracy and stability remains a challenge for forecasting researchers. This paper proposes new variational mode decomposition (VMD)-attention-based spatio-temporal network (VASTN) method that takes advantage of both temporal spatial correlations speed. First, VASTN is hybrid model combines VMD, squeeze-and-excitation (SENet), attention mechanism (AM)-based bidirectional long memory (BiLSTM). initially employs VMD to decompose the matrix into series intrinsic functions (IMF). Then, extract features at bottom model, each IMF an improved convolutional neural algorithm based on channel AM, also known as SENet. Second, it BiLSTM AM top layer aggregated capture dependencies. Finally, accumulates predictions obtain predicted reduce randomness instability original data before employing improve through mapping weight parameter learning. Experimental results real-world demonstrate VASTN’s superiority over previous related algorithms.