표 5 GBM 학습모형 패러미터

패러미터	값
Objective	Regression
Max depth	-1
Min data in leaf	5
Number of leaves	511
Number of iterations	10,000
Early stopping rounds	30
Verbosity	-1
Learning rate	0.1
Split rate (train-validation-test set)	60-20-20(%) Randomly split

주 : GBM, gradient boosting model.