Table 5.

Candidate parameters for grid search and fine-tune of parameters.

Algorithms	Candidate parameters
Logistic regression	C: [0.01, 0.1, 1, 10] class_weight: [‘balanced’]
Artificial neural network	hidden_layer_sizes: [(50, 50), (100, 100), (150, 150)]
Random forest	n_estimators: [30, 60 ,90] max_depth: [30, 60, 90] max_leaf_nodes: [30, 60 ,90] class_weight: [‘balanced’]
XGBoost	n_estimators: [3,30,60] max_depth: [3, 30] reg_lambda: [0.1, 10] gamma: [0.1, 10] min_child_weight: [0.2, 8, 20] scale_pos_weight: [3,30,300,3000,3300,3600,3900,4000,5000,6000,7000,8000] subsample:[0.3, 0.6, 1] max_delta_step:[0, 3]

These parameters came from Python machine learning package called Sklearn.