Table 4.

The optimal hyperparameters for ML models.

ML model	Optimal hyperparameters
CatBoost	iterations = 1696, learning_rate = 0.0906, depth = 4, subsample = 0.389, colsample_bylevel = 0.784, min_data_in_leaf = 10
GPR	Kernel: Constant*RBF + Constant*Matern + Constant*WhiteKernel + Constant* RationalQuadratic, gpr.alpha = 0.002
LightGBM	n_estimators = 1972, learning_rate = 0.0329, max_depth = 50, num_leaves = 10, boosting_type = Gradient Boosting Decision Tree
XGBoost	n_estimators = 1968, max_depth = 41, learning_rate = 0.0611, booster = ‘dart’, gamma = 0.01
RandomForest	random_state = 1000, n_estimators = 1134, max_depth = 22, min_samples_leaf = 2, max_features = ‘log2’, bootstrap = False
ANN	number of hidden layers = 1, neurons number of hidden layer = 12
SVR	Log(C) = 0.9988, log(epsilon) = − 74.88644548, log(gamma) = − 1.72300541