Table 5.

The selection of hyper-hyperparameters.

Algorithms Types	Hyper-parameters Values Search Range	Best Parameters
Random Forest	n_estimators: linspace(start = 50,stop = 3000,num = 60)	2800
	max_features: [‘auto’, ‘sqrt’]	auto
	max_depth: linspace(start = 10,stop = 500,num = 50)	430
	min_samples_split [2,5,10]:	2
	min_samples_leaf [1,2,4,8]:	1
Decision Tree	criterion: [‘Gini’, ‘entropy’]	Gini
	min_samples_leaf: linspace(start = 1,stop = 50,num = 10)	46
	min_impurity_decrease: linespace(start = 0,stop = 0.5,num = 20)	0.0
	max_depth [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]]:	6
BernoulliNB	Default	Default
KNeighbors	n_neighbors [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11]]:	11
	‘Metric’: [‘euclidean’,’manhattan’,’cheebyshev’,’minkowski’]	minkowski
	weights: [‘uniform’,’distance’]	distance
	p [[1], [2], [3], [4], [5], [6]]:	4
Logistic Regression	‘C’: [0.25,0.5,0.75,1]	0.5
	solver: ['newton-cg', 'lbfgs', 'liblinear', 'sag', 'saga']	newton-cg
	penalty: [L1,L2]	L2
GaussianNB	Default	Default
XGBoost	n_estimators: linespace(start = 50,stop = 3000,num = 60)	250
	max_features [1,3,5,7,9]:	1
	max_depth [[3], [4], [5], [6], [7], [8], [9], [10]]:	10
	min_samples_leaf [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]]:	9
	min_samples_split [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]]:	9
Catboost	iterations: linspace(start = 50, stop = 3000,num = 10)	600
	max_depth [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]]:	9
	subsample [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]]:	1
Lightgbm	learning_rate: [0.01,0.02,0.05,0.1,0.15]	0.1
	feature_fraction: [0.6, 0.7, 0.8, 0.9, 0.95]	0.6
	max_depth [15,20,25,30,35]:	30
	bagging_fraction: [0.6, 0.7, 0.8, 0.9, 0.95]	0.6
	lambda_l1: [0, 0.1, 0.4, 0.5, 0.6]	0
	lambda_l2: [0, 10, 15, 35, 40]	0