Table 3 The best hyperparameters of the machine learning models were obtained for data balancing with SMOTE and ADASYN

DTC

criterion

Entropy

Entropy

max_depth

None

None

min_samples_leaf

1

1

min_samples_split

2

2

MLP

activation

relu

relu

hidden_layer_sizes

(100, 50)

(100, 50)

learning_rate

constant

adaptive

solver

adam

adam

KNN

algorithm

auto

auto

leaf_size

1

1

n_neighbors

1

1

p

2

2

weights

uniform

uniform

SGDC

alpha

0.001

0.001

loss

hinge

hinge

max_iter

3000

3000

penalty

elasticnet

l2

ETC

min_samples_split

2

2

n_estimators

150

300

random_state

40

50

SVM

C

10

10

gamma

scale

auto

kernel

rbf

rbf

RFC

max_depth

None

None

min_samples_split

4

4

n_estimators

150

150

random_state

30

GB

learning_rate

0.1

0.1

max_depth

7

7

n_estimators

200

200

random_state

40

40