Prediction of Highway Tunnel Pavement Performance Based on Digital Twin and Multiple Time Series Stacking

Table 1

Hyperparameter range of each algorithm.

Algorithm name	Hyperparameters selected
XGBoost	learning_rate = [0.01, 0.1, 0.2]
	n_estimators = [100, 50]
	max_depth = [30, 50, 100]
	Gamma = [0.1, 0.2, 0.5]
	Subsample = [0.9, 0.5]
	colsample_btree = [0.9, 0.5]
RF	Criterion = [“mse,” “mae”],
	n_estimators = [10, 100]
	max_depth = [10, 50, 100]
	max_features = [0.5, 0.9]
	min_samples_split = [10]
	min_samples_leaf = [2, 10]
Ridge	Solver = {“auto,” “svd,” “cholesky,” “lsqr,” “sparse_cg,” “sag”}
	alpha = [0.1, 1, 10]
SVR	C = [0.01, 0.1, 1], kernel = [“linear,” “poly,” “rbf,” “sigmoid”]
ANN	hindden_cell_num = [32, 64, 128]
	Activation = [“tanh,” “relu”]
	Scoring = [ “neg_mean_squared_error” ]
	Epochs = [300, 100]
	batch_size = [20, 50]