Detection of epileptic seizure based on EEG signals

In this paper, the support vector machines (SVMs) is adopted for distinguishing between normal and epileptic EEG time series. The embedding dimension of electroencephalogram (EEG) time series is used as the input feature for detecting epileptic seizure automatically. Cao´s method is applied for computing the embedding dimension of normal and epileptic EEG time series. In the last work, probabilistic neural networks (PNN) was employed for detecting epileptic seizure automatically, therefore, the results obtained by SVMs are compared with those obtained by PNN in this paper. The results show that the overall accuracy as high as 100% can be achieved by both the methods; however, for the same accuracy, the experiment by SVM needs less input features than PNN.