Skip to main content
. 2021 May 3;2021:6630643. doi: 10.1155/2021/6630643

Table 2.

Results of different convolution kernel size models on the CCDD testing (%).

F1
MResDNN-37 MResDNN-57 MResDNN-77 MResDNN-79 MResDNN-711 MResDNN-357
SA 86.1 84.9 85.7 85.0 85.6 85.9
SB 93.0 93.2 92.9 93.0 93.2 93.0
TWFL 79.6 79.6 78.8 79.5 78.5 78.2
ST 95.3 95.6 95.6 95.2 95.8 95.5
CRBBB 97.5 95.1 96.6 96.2 95.6 93.8
AF 95.7 96.0 96.0 95.8 94.6 95.9
APB 61.7 60.7 58.3 58.9 59.0 58.4
I-AVB 78.6 78.4 79.1 79.2 76.6 77.6
PVC 60.4 66.6 72.8 75.3 59.0 67.3

Average
Acc 96.8 96.7 96.8 96.7 96.7 96.7
Spe 98.6 98.5 98.6 98.5 97.6 98.5
Pre 89.7 88.8 90.3 88.6 83.1 88.9
Rec 84.0 84.2 83.5 84.6 89.4 83.6
F1 86.8 86.5 86.7 86.5 86.1 86.2