Research Article
Effective Estimation of Hourly Global Solar Radiation Using Machine Learning Algorithms
Table 12
Comparison of the most successful HAGSR estimation models that were developed by using MFFNNN, KNN, M5 rules, and LibSVM for both provinces.
| Statistical indicators | Target provinces | Kahramanmaras | Isparta | MFFNN | KNN | M5 rules | LibSVM | MFFNN | KNN | M5 rules | LibSVM |
| RMSE | 0.0508 | 0.0605 | 0.0610 | 0.0675 | 0.0536 | 0.0646 | 0.0650 | 0.0752 | MAE | 0.0341 | 0.0419 | 0.0418 | 0.0501 | 0.0352 | 0.0438 | 0.0441 | 0.0573 | | 0.9656 | 0.9511 | 0.9506 | 0.9394 | 0.9488 | 0.9261 | 0.9254 | 0.8995 | SMAPE (%) | 7.79 | 8.84 | 9.01 | 12.14 | 7.77 | 8.88 | 8.42 | 12.11 | Data groups | GR2 | GR2 | GR5 | GR2 | GR3 | GR3 | GR3 | GR3 |
|
|