Short-Term Wind Speed Forecasting Based on the EEMD-GS-GRU Model

To improve the accuracy of short-term wind speed forecasting, we proposed a Gated Recurrent Unit network forecasting method, based on ensemble empirical mode decomposition and Grid Search Cross Validation parameter optimization algorithm. In this study, first, in process decomposing, set was introduced to divide time series into high-frequency modal, low-frequency trend using Pearson correlation coefficient. Second, during optimization, grid algorithm employed GRU model search for combination optimal parameters. Third, improved driven with decomposed components predict new components, which were used obtain predicted by modal reorganization. Compared other models (i.e., LSTM, GS-LSTM, EEMD-LSTM, EEMD-GS-LSTM), applied case study farm, located northwest China. The results showed that presented could reduce error (RMSE) from 1.411 m/s 0.685 can forecasts. This provides approach forecasting.