Recognition of epilepsy from non-seizure electroencephalogram using combination of linear SVM and time domain attributes

Classification of neural disorders, like epilepsy, can be performed efficiently using soft computing methods. Previously many methods of detecting epilepsy using various time and time-frequency domain features have been proposed. Our study proposes a unique feature set, comprising of time domain features, Waveform Length, Root Mean Square, Mean Absolute Value and Zero Crossing, combining them with Linear Support Vector Machine to classify a set of EEG signals into epileptic and non-epileptic under non-seizure condition. Our proposed classification approach yields an accuracy of 95%.