Singular Values as a Detector of Epileptic Seizures in EEG Signals

This paper introduces a new method based on the Singular Values of EEG signals for the detection of epileptic seizures. Singular Value Decomposition was performed on an EEG signal in epochs of 8 seconds and Singular Values were extracted from each epoch. These singular values were fed into Support Vector Machine (SVM) for a binary classification between epileptic seizure and non- seizure events. Singular Values of EEG signals proved to be a very good feature for the detection of epileptic seizures and gave a classification accuracy of 90%, and an average sensitivity and specificity of 91% and 89%, respectively.