Table 3 Detection accuracies of LDA, modified KNN and LGBM classification algorithms at the time of the CS event for all subtracted SNNr cubes
Detection CS onset | ||||||
|---|---|---|---|---|---|---|
I–O connection | 32 Channels trained | 5 Channels trained | ||||
LDA | KNN | LGBM | LDA | KNN | LGBM | |
1471 reservoir + I–O | 57.8% | 70.30% | 75.0% | 56.3% | 57.80% | 65.6% |
32 | 65.6% | 75.00% | 70.3% | N/A | N/A | N/A |
5 | 50.0% | 62.50% | 67.2% | 53.1% | 62.50% | 60.1% |
FC6 | 59.4% | 65.60% | 62.5% | 53.1% | 67.20% | 65.6% |
Fp2 | 42.2% | 56.30% | 59.4% | 43.8% | 59.40% | 59.4% |
Fp1 | 46.9% | 60.90% | 57.8% | 0.00% | 64.10% | 59.4% |
Cz | 25.0% | 56.30% | 54.7% | 53.1% | 56.0% | 64.1% |
O2 | 51.6% | 53.10% | 45.3% | 0.00% | 50.0% | 59.4% |
Best combo out of 5 | FC6 59.4% | Fp2, Cz 68.80% | Fp1,Cz 68.8% | FC6 + Fp2 + Cz 57.8% | Fp2, Cz 68.80% | FC6,Fp1,Cz 68.8% |