Skip to main content

Table 3 The results of evaluation of machine learning and deep neural network models for determining age group

From: GADNN: a revolutionary hybrid deep learning neural network for age and sex determination utilizing cone beam computed tomography images of maxillary and frontal sinuses

Method

Age Groups

Precision

Recall

F1-Score

Accuracy

Logistic Regression with SMOTEa

a

0.50

0.56

0.53

0.36

b

0.28

0.31

0.29

c

0.29

0.10

0.15

d

0.33

0.50

0.40

Multi-layer Perceptron with SMOTE

a

0.48

0.61

0.54

0.38

b

0.27

0.25

0.26

c

0.29

0.10

0.15

d

0.37

0.56

0.44

Random Forest with SMOTE

a

0.65

0.83

0.73

0.57

b

0.60

0.56

0.58

c

0.38

0.25

0.30

d

0.57

0.67

0.62

Deep Learning without SMOTE

a

0.40

1.00

0.57

0.37

b

0.25

0.33

0.29

c

0.57

0.32

0.41

d

0.14

0.12

0.13

Deep Learning with SMOTE

a

0.59

0.87

0.70

0.62

b

0.50

0.47

0.48

c

0.46

0.38

0.41

d

0.84

0.75

0.80

GADNNb with SMOTE

a

0.76

0.89

0.82

0.68

b

0.67

0.67

0.67

c

0.44

0.50

0.47

d

0.83

0.65

0.73

  1. athe synthetic minority oversampling technique; bGenetic algorithm based deep neural network model