Table 3 AUC of ROC and PR curves of each method on two benchmark datasets
From: A multi-context learning approach for EEG epileptic seizure detection
| Â | CHB-MIT Dataset | Bonn Dataset | ||
|---|---|---|---|---|
Method | AUC-ROC | AUC-PR | AUC-ROC | AUC-PR |
PSVM | 0.7283 | 0.6012 | 0.6987 | 0.8111 |
PNN | 0.7754 | 0.5487 | 0.8548 | 0.8608 |
PSDAEs | 0.7166 | 0.3679 | 0.9608 | 0.9620 |
Context-EEG [17] | 0.8784 | 0.8304 | 0.9930 | 0.9925 |
WT-PSVM | 0.8552 | 0.7411 | 0.9706 | 0.8603 |
WT-PNN | 0.9432 | 0.7701 | 0.9696 | 0.9718 |
WT-PSDAEs | 0.8214 | 0.7577 | 0.9779 | 0.9853 |
WT-CtxEEG | 0.9782 | 0.9249 | 0.9994 | 0.9974 |
WT-CtxFusionEEG | 0.9874 | 0.9649 | 1.0000 | 0.9980 |