Probabilistic Evaluation of Hydraulic Fracture Performance Using Ensemble Machine Learning

Table 3

Optimal hyperparameter values of the GBDT.

Hyperparameter	Name	Ranges	Optimal values
No. of trees	N_estimators	1~200	110
Max. depth	max_depth	2~8	3
No. of features to split	max_features	sqrt; auto	sqrt
Min. number of samples	min_samples_split	2~12	4
Learning rate	learning_rate	0.08~0.13	0.11