Detection Of Epilepsy Using Single Channel Eeg Features

Abstract

The nervous disorder that results in abnormal electrical activity (seizure) is understood as encephalopathy. The diagnosing and treatment of those seizures area unit difficult for non-clinicians further as for neurophysiologists, motivates the young researchers to develop automatic seizure detection system. The foremost necessary section whereas process electroencephalogram input for the extraction of features and classification. In this paper, the options of healthy and ictic subjects with high order applied math options like variance (STD), Power, Skewness, and Kurtosis area unit analyzed. Several analyses had been done before with a similar techniques, however the options and methodology used area unit totally different during this paper. Aclassifies performance have been evaluated based on the confusion matrix within which ANN outperforms a straightforward two-class SVM and KNN.