Table 2 Hyperparameters, along with the best values, were tested for LSTM, Bi-LSTM, GRU, ANN-MLP, and XGBoost models.

XGBoost

n_estimators

[50, 100, 150]

Best: 100

learning_rate

[0.01, 0.1, 0.2]

Best: 0.1

max_depth

[3, 5, 7]

Best: 5

min_child_weight

[1, 3, 5]

Best: 1

LSTM

neurons

[50, 100]

Best: 50

epochs

[50, 100]

Best: 50

batch_size

[32, 64]

Best: 32

Bi-LSTM

neurons

[50, 100]

Best: 50

epochs

[50, 100]

Best: 50

batch_size

[32, 64]

Best: 32

GRU

neurons

[50, 100]

Best: 50

epochs

[50, 100]

Best: 50

batch_size

[32, 64]

Best: 32

MLP

neurons

[50, 100]

Best: 50

epochs

[50, 100]

Best: 50

batch_size

[32, 64]

Best: 32