Table 3 Evaluation metrics on training set ordered by increasing R² (Lower to higher Efficiency).
From: Comparative analysis of deep learning architectures in solar power prediction
Model | RMSE | MAE | MAPE | R² |
|---|---|---|---|---|
Transformer | 862.0321 | 745.9043 | 148581.6866 | 0.1551 |
LSTM | 525.9054 | 418.8194 | 25620.8813 | 0.6855 |
GRU | 486.6311 | 385.2489 | 24411.0665 | 0.7307 |
CNN | 462.4756 | 345.9687 | 18104.5631 | 0.7568 |
InformerLite | 423.7508 | 313.7759 | 30930.4242 | 0.7958 |
SimpleRNN | 410.9217 | 315.5270 | 21515.0806 | 0.8080 |
TCN | 378.1977 | 270.5365 | 11125.2028 | 0.8374 |
Autoencoder | 341.0466 | 234.7443 | 13257.2408 | 0.8677 |
