Table 5. The per-segment SA detection and per-recording classification performance in the UCDDB database.
Classifier | Per-segment | Per-recording | |||||
---|---|---|---|---|---|---|---|
Accuracy (%) | Sensitivity (%) | Specificity (%) | Accuracy (%) | Sensitivity (%) | Specificity (%) | Corr. | |
SVM | 70.6 | 32.7 | 83.3 | 92.3 | 100.0 | 50.0 | 0.251 |
LR | 69.6 | 34.7 | 81.3 | 84.6 | 90.9 | 50.0 | 0.107 |
KNN | 66.1 | 38.1 | 75.4 | 84.6 | 100.0 | 0.0 | 0.373 |
MLP | 67.2 | 38.5 | 76.8 | 92.3 | 100.0 | 50.0 | 0.263 |
LeNet-5 | 71.8 | 26.6 | 86.9 | 92.3 | 90.9 | 100.0 | 0.624 |