A New Approach in Epilepsy Diagnosis using Discrete Wavelet Transformation and Analysis of Variance

Epilepsy is a chronic disorder and outbreak of brain function, caused by the abnormal and intermittent electric discharge of brain neurons. Electroencephalogram signals represent brain activities, and one of the methods of diagnosing epilepsy is using EEG brain signals. In this article, a new method for diagnosing epilepsy using EEG signal processing is presented. At first, the EEG signal is divided into five frequency sub-bands using Discrete Wavelet Transformation (DWT). Then, the features are extracted from five frequency sub-bands, and the best features are selected by the analysis of variance (ANOVA) method. Finally, by using the Support Vector Machine (SVM) algorithm, these features are used to classify seizure and non-seizure EEG signals. The simulation results from the Bonn university dataset affirm the suggested approach's advantage in comparison with some other basic classical methods in terms of accuracy, sensitivity, and specificit.

A New Approach in Epilepsy Diagnosis using Discrete Wavelet Transformation and Analysis of Variance Keywords:

Epileptic Seizure , Features selection , Electroencephalogram signals , Support vector machine (SVM)