Classification of epileptic EEG based on improved empirical wavelet transform

Abstract Electroencephalography (EEG) is the most commonly used method in diagnosis of epilepsy diseases. In order to identify EEG signals more effectively, an automatic identification based on improved empirical wavelet transform (EWT) proposed. Firstly, view difficulty spectral division signal processing by transform, improvement measure proposed, that is, average difference spectrum obtained replace and then a number component with epileptic characteristic can be from original signal. Afterward, feature extraction classification are completed through common spatial pattern AdaBoost algorithm. Simulation analysis was carried out Bonn data set, healthy people patients were identified classified interictal ictal periods, high accuracy achieved.