An Approach to Identify Different Types of EEG Epileptic Signals Based on Higher-Order Spectra (HOS) Features

The aim of the research consists in finding an efficient features vector in order to classify the type of seizures of epileptic patients. The handled database consists of focal and nonfocal EEG recordings from epileptic patients. It is performed higher-order spectral analysis applying bispectrum and bicoherence methods to accurately identify the type of seizure so that epileptic patients could undergo surgical rejection of epileptic area. This approach is based on a new way of forming the features vectors, from 10% bispectrum and 90% bicoherence. In the classification stage, the k-Nearest neighbors (kNN) classifier was used, because it performs the best, leading to a maximum value of the classification rate of 99.55%, to a sensitivity of 100%, and a specificity of 99.09%.