Skip to main content

Table 2 Comparison results of different features in 1D CNNs

From: Feature extraction and classification of heart sound using 1D convolutional neural networks

Categories of features

Convolutional kernel shape

Accuracy rate (%)

F-score (%)

MFCC

1×13

85.64

86.23

Deep feature

 

95.68

96.30

MFCC

1×26

90.23

91.01

Deep feature

 

97.63

98.65

MFCC

1×39

91.02

92.78

Deep feature

 

97.85

98.67

MFCC

1×52

89.36

89.88

Deep feature

 

95.21

96.73

MFCC

1×65

93.21

94.01

Deep feature

 

93.87

95.23