Table 5.
All given accuracies are the result of the Monte Carlo Technique.
| Feature vector | Power spectra features | Theta band features | ||
|---|---|---|---|---|
| Validation technique | Leave-one-out | Direct classification | Leave-one-out | Direct classification |
| EEG1 | 68% | 64% | 79% | 63% |
| EEG2 | 78% | 77% | 93% | 82. 5% |
In the table, it is seen that EEG2 gives better results than EEG1 for all cases. Feature selection technique also improves the results. The maximum accuracies for each category are in bold.