Table 2 Improvement in the predictive performance of PBNN with MI-based feature selection.
From: Parallel boosting neural network with mutual information for day-ahead solar irradiance forecasting
Location | Models | RMSE | MAE | MSE | MAPE | NRMSE |
|---|---|---|---|---|---|---|
(\(W~m^{-2}\)) | (\(W~m^{-2}\)) | (\(W~m^{-2})^2\) | (%) | (%) | ||
Islamabad | Selected features | 14.06 | 8.36 | 197.77 | 0.26 | 0.29 |
All features | 25.18 | 17.39 | 633.97 | 0.49 | 0.52 | |
San Diego | Selected features | 17.23 | 5.26 | 296.84 | 0.12 | 0.32 |
All features | 36.74 | 21.86 | 1349.92 | 0.46 | 0.7 |
