Table 2 Prediction accuracies of LDA, modified KNN and LGBM classification algorithms at baseline 90–92 s for all subtracted SNNr cubes
Prediction 90–92 s | ||||||
|---|---|---|---|---|---|---|
I–O connection | 32 Channels trained | 5 Channels trained | ||||
LDA | KNN | LGBM | LDA | KNN | LGBM | |
1471 reservoir + I–O | 53.1% | 67.20% | 60.9% | 50.0% | 65.60% | 73.4% |
32 | 57.8% | 70.30% | 64.1% | N/A | N/A | N/A |
5 | 51.6% | 68.80% | 70.3% | 54.7% | 64.10% | 68.8% |
T8 | 48.4% | 71.90% | 60.9% | 50.0% | 62.50% | 56.3% |
CP6 | 59.4% | 54.70% | 56.3% | 56.3% | 64.10% | 67.2% |
Fz | 50.0% | 60.90% | 59.4% | 0.00% | 64.10% | 54.7% |
FC5 | 18.8% | 60.90% | 56.3% | 42.2% | 60.90% | 54.7% |
T7 | 23.4% | 65.60% | 53.1% | 42.2% | 57.80% | 54.7% |
Best combo out of 5 | T8 + CP6 + Fz 59.4% | T8 73.40% | T8 61.3% | T8,CP6,Fz,FC5 57.8% | T8, CP6 75% | T8,CP6 66.1% |