Skip to main content
. 2023 Jan 26;23(3):1365. doi: 10.3390/s23031365

Table 4.

Cloud training and testing classification results: accuracy (ACC), precision (PREC), sensitivity (SEN), and F1 score metrics (%) of the proposed model variants. Model variations include: the number of classes: {3,4,5}, Conv1D trainable parameter: {True, False, Default}, MF kernel size NK: {32, 64}, the input feature to the Conv1D layer: {Signal, Derivative, Both}, and the class weight training parameter: {SET, Not SET}. Best results per row (metric) are highlighted in red.

Model ID Number 1 2 3 4 5 6 7 8 9
Model Variations Number of Classes 3 3 3 3 3 3 3 4 5
Conv1D Trainable DEF TRUE FALSE DEF TRUE TRUE TRUE TRUE TRUE
Layer Kernel Size NK 64 32 32 32 32 32 64 32 32
Input Feature DREV DREV DREV BOTH SIG DREV DREV DREV DREV
Class Weight Param SET SET SET SET SET NOT NOT NOT NOT
Results Total Number of Params 1619 1267 1267 1619 1267 1267 1619 1352 1420
Number of Trainable Params 1597 1245 879 1597 1245 1245 1597 1330 1389
Training Results % Training Time (s) 118.26 94.40 90.61 146.337 122.67 83.76 96.95 77.90 103.76
Model Accuracy 99.30 98.61 93.00 96.04 96.17 99.31 99.42 98.59 98.94
Average F1-Score 95.35 91.54 75.57 83.34 82.18 94.71 95.55 86.68 72.46
Testing Results % Model Accuracy 96.48 95.58 94.42 94.78 92.90 98.18 97.94 97.06 96.93
Normal Class—N PREC 99.38 99.47 99.39 99.57 99.67 99.00 99.06 97.73 97.96
SEN 96.96 95.86 94.67 94.91 92.98 99.10 98.31 99.14 98.85
F1 98.16 97.63 96.98 97.18 96.21 99.05 98.68 98.43 98.40
Supraventricular Class—SVEB PREC 58.06 51.49 43.85 56.09 39.19 82.68 69.72 84.29 90.78
SEN 85.30 88.51 95.59 92.54 88.13 81.60 81.33 79.42 74.52
F1 69.09 65.11 60.12 69.84 54.26 82.14 75.08 81.78 81.85
Ventricular Class—VEB PREC 90.20 87.93 92.23 74.10 78.07 95.63 94.53 94.99 93.39
SEN 96.34 95.65 90.31 94.32 94.47 95.00 95.47 90.13 95.13
F1 93.17 91.63 91.26 82.99 85.49 95.31 95.00 92.50 94.25
Average Scores PREC 82.55 79.63 78.49 76.59 72.31 92.44 78.77 69.25 56.42
SEN 92.87 93.34 93.53 93.92 91.86 91.90 91.70 67.17 53.70
F1 86.81 84.79 82.79 83.34 78.65 92.17 89.59 68.18 54.90