EEG Based Brain State Classification Technique Using Support Vector Machine -A Design Approach

Brain Computer Interface is a good beneficial route for severely physically challenged person who is underprivileged to communicate in conventional way or have lost their ability to speak. The cause of the work carried in the paper is to solve the problem of patients suffering from neurological disorder and its disabilities that give rise to this research. In the proposed work Electroencephalogram (EEG) based brain state signal measurement method is use to record the brain activity which is source of communication system between patient and outside world. Electroencephalogram is a non-muscular channel between the human brain and a computer system is provided by brain-computer interface (BCI) in which electrical activity are recorded for perusal of EEG signals by the brain. This signals are then decomposed into smaller segments of signal by Time frequency approaches (T-F) like fast Fourier transform & short time Fourier transform. Both these techniques acts as a feature extraction method followed by training of the data and the classification is done by using support vector machine. The performance parameters like accuracy, precision, sensitivity, specificity are calculated based on the values of evaluation metrics and overall system accuracy comes out to be 92%. The four classified signals can be used as Communication messages by the patients which will help to solve the speech impairment problem of disabled person.