This paper provides a theoretical analysis of modern forecasting methodologies for power generation in hybrid wind–solar systems. It discusses statistical, physical, and machine learning approaches, with emphasis on hybrid architectures integrating numerical weather prediction (NWP) data with deep learning models. Evaluation metrics and methodological frameworks are analyzed, highlighting their applicability to hybrid energy forecasting.

hybrid renewable energy, forecasting, wind–solar system, LSTM, Transformer, ARIMA, NWP