| | Ref. | Used models | Data division | Best model | Statistical matrices |
| | [1] | Random forest, XGBoost, and deep learning | 70% for training;
30% for testing | XGBoost | R, MAE, RMSE |
| | [43] | Random forest | 10-fold cross-validation | RF | , RPE |
| | [52] | SVM, ANN | 80% for training;
20% for testing | ANN | R, MSE |
| | [53] | EEMD-GRNN
ANFIS
MLR
MLP | 90% for training;
10% for testing | ANFIS | R, , MAE, RMSE |
| | [54] | GTWR
LR
ANN
ANIFS
GRNN | — | GRNN. | R, MAE, RMSE |
| | [55] |
| 10-fold cross-validation | ANN | R, MAE, MAPE, RMSE, IA |
| | [56] | EEMD-GRNN
GRNN
MLR
PCR
ARIMA | 90% for training;
10% for testing | EEMD- GRNN | MAE, MAPE, RMSE, IA |
| | [57] | MLP | 70% for training;
15% for validation;
15% for testing | MLP | MAE, RMSE, IA |
| | [58] | XGBoost includes NELRM | 10-folds cross-validation | XGBoost. | , RMSE, MAPE |
| | [59] | RF model | 10-folds cross-validation | RF | , RMSPE, MPE |
|
|
