Table 5 Best proposed forecasting models vs. default models based on SVR and ANN
From: Short-term photovoltaic power production forecasting based on novel hybrid data-driven models
| Â | \(CSO-SV{R}_{RB}\) | \(SV{R}_{RB}^{D}\) | \(PSO-SV{R}_{linear}\) | \(SV{R}_{linear}^{D}\) | \(SSO-BPN{N}^{1}\) | \(CSO-BPN{N}^{2}\) | \(BPN{N}^{D}\) |
|---|---|---|---|---|---|---|---|
\(RMSE \; (\text{kW})\) | 4.4795 | 8.9733 | 9.1334 | 9.9392 | 4.8460 | 4.5692 | 5.2289 |
\(nRMSE \; (\%)\) | 4.3417 | 8.6972 | 8.8524 | 9.6334 | 4.6969 | 4.4286 | 5.0681 |
\(MAE \; (\text{kW})\) | 2.5661 | 5.2218 | 4.8617 | 4.9399 | 3.1713 | 2.9614 | 3.2892 |
\(nMAE \; (\%)\) | 2.4871 | 5.0612 | 4.7121 | 4.7879 | 3.0738 | 2.8703 | 3.1879 |