Automatic Epileptic Seizure Classification using MODWT and SVM

Abstract

Epilepsy is the neurological disorder that makes tough for epileptic patients to survive a natural life. Because classifying the electroencephalography (EEG) records is a difficult job for the clinician. This study presents effective epileptic seizure detection based on Maximal overlap discrete wavelet transform (MODWT) is proposed for the decomposition of EEG signals. The important features are computed from the decomposed wavelet coefficient based on its statistical measures such as mean and standard deviation. The extracted statistical measures are then classified by using support vector machine (SVM) classifier. In this study provides the classification between normal and focal EEG signals and also between normal and generalized EEG signal. The proposed method is verified using karunya EEG database and it produces the most promising classification performance with an overall accuracy of 95.8%, sensitivity of 92.30%, and specificity of 100 % for the classification of normal and focal EEG signals. Also it provides an accuracy of 91.7%, sensitivity of 85.71%, and specificity of 100 % for the classification of normal and generalized EEG signals.