Title: Adaptive combination methods of autoregressive parameters for epileptic EEG signals classification
Authors: Boukari Nassim
Addresses: Department of Electrical and Electronics Engineering, Faculty of Engineering, Annaba University, Annaba 23000, Algeria
Abstract: Epilepsy, one of the most common neurological diseases, affects over 50 million people worldwide. Epilepsy can have a broad spectrum of debilitating medical and social consequences. This paper illustrates the use of adaptive combination autoregressive parameters for the feature extraction. The multilayer perceptron neural network is selected for the classification of electroencephalogram signals (EEG). Five types of EEG signals (Normal (A, B), Interictal (C, D), and Ictal (E) from Bonn University) were classified with the accuracy of 97.66% by the adaptive combination autoregressive parameters.
Keywords: neural networks; electroencephalograms; EEG signals; signal classification; autoregressive coefficients; adaptive weights combination; epileptics; epilepsy; feature extraction.
International Journal of Biomedical Engineering and Technology, 2016 Vol.22 No.1, pp.47 - 57
Received: 26 Oct 2015
Accepted: 22 Dec 2015
Published online: 08 Sep 2016 *