Classification of EEG signals using fractional calculus and wavelet support vector machine

Abstract

The behavior of many physical and biological processes and systems can be described satisfactorily by fractional order models. A new method, termed fractional linear prediction (FLP) based on fractional calculus, is used to model ictal and seizure-free EEG signals. Through numerical simulations it is demonstrated that, the EEG signal can be modeled accurately, by using a few integrals of fractional orders as basis functions. The parameters obtained from modeling are used for analysis and classification using support vector machines (SVM). It is found that improvements in classification accuracy is possible by using wavelet support vector machines using wavelet kernel functions such as Mexican hat wavelet and Morlet wavelet.