Table 3.

Selection of model parameters.

	Parameter	Explanation	Value
CART	Max depth	maximum depth of decision tree	12
	Min samples split	minimum number of samples required for subdividing internal nodes	8
	Min samples leaf	minimum number of samples for leaf nodes	3
ABR frame	Base estimator	weak regression learner	CART
	Loss	loss function, there are three choices of linear, square and exponential	exponential
	N estimators	maximum number of iterations of the weak learner	80
	Learning rate	the step size of the update parameter, too small will slow down the iteration speed	0.001