Table 6 Results of hyper-parameter optimization for machine learning models.

Extra Tree

{‘et__bootstrap’: False, ‘et__max_depth’: 15, ‘et__min_samples_leaf’: 4, ‘et__min_samples_split’: 2, ‘et__n_estimators’: 300}

88.04

92.30

88

88

88

Random Forest

{‘rf__max_depth’: 10, ‘rf__min_samples_split’: 10, ‘rf__n_estimators’: 200}

87.50

92.57

88

88

88

Ada Boost

{‘ada__algorithm’: ‘SAMME’, ‘ada__learning_rate’: 1, ‘ada__n_estimators’: 100}

85.87

92.40

87

86

86

Gradient Boosting

{‘gb__learning_rate’: 0.1, ‘gb__max_depth’: 3, ‘gb__n_estimators’: 100}

89.22

94.31

89

89

89

CatBoost Before Fine-tunning

(iterations = 500, learning_rate = 0.03, depth = 8, l2_leaf_reg = 5, bagging_temperature = 1, border_count = 128, random_strength = 1, od_type = ’Iter’, od_wait = 50, verbose = 0, random_state = 42)

88

88

88

88

88

CatBoost after Fine-tunning

(iterations = 200, learning_rate = 0.1, depth = 6,verbose = 0)

99.02

95

99.04

99.02

99.02