Table A6.

Hyperparamter bounds for Bayesian optimization.

XGBoost Hyper Parameters	Lower Bound	Upper Bound
scale_pos_weight	Y = 0/Y = 1	2 + Y = 0/Y = 1
n_estimators	10	1000
learning_rate	0.1	1
min_child_weight	1	10
max_depth	3	12
subsample	0	1
colsample_bytree	0.3	1
gamma	0	5
reg_alpha	1 × 10−5	0.75
reg_lambda	1 × 10−5	0.45

Y = 0: patient without relapse at 18 months, Y = 1 patients with relapse at 18 months. Bold: highlight labels of lines.