Table 5. The classification accuracies obtained for five different approaches in the classification of multi-class EEG signals with 5-FCV.

The used method	The obtained Classification accuracy (%)
Alone Random Forests classifier with raw EEG signals	64.80
The first approach (the combination of 125-TDM features and Random Forests classifier)	71.90
The second approach (the combination of OVO and Random Forests classifier)	91.08
The third approach (the combination of OVA and Random Forests classifier)	89.00
The fourth approach (the combination of OVA, ADASYN and Random Forests classifier)	91.72