Hybrid Models for Daily Global Solar Radiation Assessment

Daily solar radiation forecasting has recently become critical in developing energy and its integration into grid systems. Despite the huge number of proposed techniques, an accurate estimation remains a significant challenge because non-stationary variation components due to continuously changing climatic conditions. Usually, several input data predictors are used for process, which can cause redundancy correlation between features. This work assesses set feature selection techniques check their ability select relevant reduce redundant irrelevant information. An Artificial Neural Network is fit measured based on selected The developed model evaluated through various objective evaluation metrics using historical three years at Ghardaiaregion inAlgeria. Results show effectiveness method, where values 0,0189, 0.0286, 5.4387, 98.28% have been found as MABE, RMSE, nRMSE r, respectively.