| Model type | Forecasting model | Hyperparameter | Range | Best hyperparameter |
| Proposed model | CNN-LSTM (2 convolutional layers with 64 filters, kernel size 3, 1 max pooling layer with size 1, 1 dropout layer, 1 LSTM layer with 200 units, 1 fully connected layer) | Epochs | (32, 1000) | 472 | Batch_size | (2, 30) | 22 | Verbose | (0, 1) | 1 |
| Deep learning model | CNN (2 convolutional layers with 64 filters, kernel size 3, 1 max pooling layer with size 1, 1 fully connected layer) | Epochs | (32, 1000) | 472 | batch_size | (2, 30) | 22 | Verbose | (0, 1) | 1 | LSTM (1 LSTM layer with 200 units) | Epochs | (32, 1000) | 472 | Batch_size | (2, 30) | 22 | Verbose | (0, 1) | 1 |
| Statistical model | ARIMA | p | (0, 10) | 9 | d | (0, 3) | 2 | q | (0, 3) | 2 | FBProphet | Changepoint_prior_scale | (0.0001, 0.5) | 0.5 | Seasonality_prior_scale | (0.01, 10) | 0.25 | Seasonality_mode | (0, 1) | 1 |
| Linear model | LR | Fit_intercept | [True, false] | True | n_jobs | (ā1, 1) | ā1 | Ridge | Alpha | (1, 5) | 5 | Lasso | Alpha | (1, 5) | 5 |
| Ensemble model | XGBoostR | n_estimators | (0, 1000) | 545 | Max_depth | (0, 25) | 6 | Reg_alpha | (0, 5) | 1 | Reg_lambda | (0, 5) | 3 | Gamma | (0, 5) | 1 | Learning_rate | (0.005, 0.5) | 0.1225 | AdaBoostR | n_estimators | (0, 1000) | 545 | RFR | n_estimators | (0, 1000) | 545 | GBR | n_estimators | (0, 1000) | 545 | ETR | n_estimators | (0, 1000) | 545 | BaggingR | n_estimators | (0, 1000) | 545 |
| Machine-learning model | GPR | Kernel | DotProduct, Matern, RBF, WhiteKernel | DotProduct | Alpha | (0, 1) | 0.16000000000000003 | SVR | Kernel | rbf, poly | poly | C | (0, 10) | 1.5 | Gamma | (0, 5) | 3 | Epsilon | (0, 1) | 0.1 | DTR | Max_depth | (0, 25) | 5 | KNNR | n_neighbors | (0, 10) | 3 |
|
|