Research on Wind Power Prediction Based on Deep Learning

Wind power prediction is not only vital for the secure and stable operation of power grids but also significantly supports the economic benefits, environmental protection, and policy-making of wind farms. Wind power prediction currently confronts challenges such as substantial data demands, stringent data quality requirements, limited interpretability of prediction models, and a high degree of model overfitting. This paper introduces a deep learning time-series prediction model based on stacking. By comparing seven individual deep learning temporal prediction models, this approach demonstrates higher prediction accuracy than conventional deep learning temporal models.