Table 7 The classification accuracy based on subsets with 1–10 channels selected by FA algorithm.

From: Detection of Parkinson’s disease from EEG signals using discrete wavelet transform, different entropy measures, and machine learning techniques

No. of selected channels by FA

#Feature/vector

FE Method + classifier

DWT + LogEn + KNN

DWT + LogEn + SVM

DWT + ThEn + KNN

DWT + SuEn + KNN

DWT + TShEn + KNN

1

5

88.29 ± 4.31

76.00 ± 5.36

80.24 ± 4.69

79.90 ± 4.75

79.04 ± 4.60

2

10

94.93 ± 2.82

91.82 ± 3.60

89.97 ± 3.86

91.52 ± 3.27

92.05 ± 3.36

3

15

96.63 ± 2.15

96.62 ± 2.50

95.15 ± 3.02

95.85 ± 2.36

96.54 ± 2.23

4

20

97.70 ± 1.90

96.86 ± 1.96

97.33 ± 2.14

97.39 ± 1.94

97.74 ± 2.04

5

25

97.76 ± 1.69

97.29 ± 2.11

98.27 ± 1.79

98.22 ± 1.57

98.05 ± 1.92

6

30

–

98.46 ± 1.59

98.70 ± 1.79

98.83 ± 1.48

98.58 ± 1.61

7

35

98.27 ± 1.93

98.98 ± 1.37

99.01 ± 1.17

99.13 ± 1.23

98.70 ± 1.53

8

40

98.19 ± 1.68

–

99.21 ± 1.05

99.44 ± 1.02

99.19 ± 1.09

9

45

–

–

–

99.47 ± 1.03

–

10

50

–

99.34 ± 1.12

99.44 ± 0.87

–

99.72 ± 0.70

Back to article page