An Efficient EEG Channels-Selection Approaches For Epilepsy Seizure Prediction

In this study, we are interested in the epilepsy seizures problem. Indeed, used binary SVM to predict ongoing and multiclass different states of patients' epilepsy. Brain activity is as an efficient source for predicting seizures, it's recorded Electroencephalography (EEG) segments signal. We propose compare paper, three ideas select channels: highest frequency channels, channels left part head, right head. A features extraction stage important produce a rich relevant dataset, effect, 22 calculated each segment 5 min from EEG named pre-ictal, one-versus-all multi-class four classes (pre-ictal, ictal, inter-ictal, post-ictal). classification rate toward 97%, on selected corpus, was achieved by (binary multiclass) with majority patients.