Table 5 Confusion matrices of various machine learning (ML) and multi-layer perceptron (MLP) models

From: Non-contact lung disease classification via orthogonal frequency division multiplexing-based passive 6G integrated sensing and communication

Algorithms

Actual/Guess

AST

COPD

ILD

PN

TB

N

SVM (Quad.)

AST

83.4

1.1

9.1

2.9

1.5

2.0

COPD

0.9

80.6

8.9

2.7

2.9

3.9

ILD

0.8

0.1

95.0

2.1

0.3

1.7

PN

1.4

1.0

9.1

84.1

2.0

2.4

TB

0.9

1.1

8.8

3.1

82.4

3.6

N

0.5

0.5

6.1

2.9

1.0

88.9

SVM (Cubic)

AST

84.1

2.0

3.4

5.0

2.1

3.4

COPD

2.0

81.5

2.6

5.9

3.9

4.1

ILD

2.8

0.3

88.9

6.2

0.9

0.8

PN

3.4

1.4

5.2

83.5

2.2

4.4

TB

2.3

2.0

4.0

6.4

81.4

3.9

N

1.3

0.2

3.5

5.4

0.7

88.8

MLP (Narrow)

AST

92.5

1.9

0.5

1.1

2.7

1.3

COPD

0.7

95.3

0.8

0.5

1.3

1.3

ILD

0.7

0.4

94.1

1.1

1.6

2.1

PN

1.3

0.9

1.3

92.7

1.7

2.0

TB

1.2

1.9

1.3

1.2

92.2

2.1

N

0.9

1.2

1.9

1.5

2.0

92.6

MLP (Medium)

AST

96.8

0.8

0.5

0.2

1.1

0.5

COPD

0.4

96.7

0.7

0.1

1.0

1.1

ILD

0.2

0.6

96.6

0.5

1.1

0.9

PN

0.6

0.6

0.8

95.7

1.1

1.2

TB

1.3

1.0

0.8

1.0

94.7

1.2

N

0.5

0.8

0.9

1.1

2.0

94.6

MLP (Wide)

AST

95.7

0.5

0.6

0.7

0.8

1.6

COPD

0.7

95.8

0.3

0.2

1.6

1.4

ILD

0.4

0.4

97.5

0.2

0.6

0.9

PN

0.7

0.3

0.5

95.8

1.3

1.3

TB

0.7

1.6

1.1

0.8

92.7

3.1

TB

0.7

1.6

1.1

0.8

92.7

3.1

MLP (Bilayered)

AST

92.9

1.3

0.8

0.6

2.3

2.0

COPD

1.0

93.6

0.3

1.6

1.3

2.1

ILD

1.5

1.2

92.2

1.3

2.0

1.8

PN

2.1

2.9

1.2

91.2

1.5

1.1

TB

3.1

1.2

1.2

1.2

91.8

1.5

N

1.6

1.4

0.9

0.9

2.0

93.2

MLP (Trilayered)

AST

90.7

1.6

2.4

1.6

2.8

0.9

ILD

3.3

1.6

89.3

1.7

2.5

1.6

PN

3.2

2.3

3.6

86.9

3.0

1.1

TB

3.3

2.7

2.3

3.0

86.6

2.1

N

2.0

1.9

1.7

1.6

3.4

89.4

  1. The diagonal values (in bold) represent the number of correctly classified instances for each class.
  2. AST Asthma, N Normal.