Investigation on performance analysis of Hybrid classifiers hmm neural networks Extreme learning machines and FPGA Implementation of WNN for classification of Epilepsy risk levels from EEG signals

Abstract

Epilepsy is a chronic neurological disorder of the brain newlineapproximately 1 of the world population suffers from epilepsy Epilepsy is newlinecharacterized by recurrent seizures that cause rapid but revertible changes in newlinethe brain functions Temporary electrical interference of the brain roots newlineepileptic seizures The occurrence of an epileptic seizure appears newlineunpredictable A behavioral seizure is the clinical manifestation of epilepsy newlineas perceived by the patient and seen by the observer An electrographic EEG newlineseizure is defined as an abnormal bursting of EEG patterns Classification of newlineepilepsy risk levels according to International Standard is not easy because newlineindividual laboratory findings and symptoms are often unconvincing The newlineEEG signal is commonly used as a diagnostic indicator for investigating brain newlineactivities under different physiological conditions newlineThis research work investigates to analyze the dimensionality newlinereduction techniques and hybrid classifiers for a quick classification of newlineepilepsy risk levels from EEG Signals The dimensional reduction is carried newlineout by the Singular Value Decomposition SVD Principal Component newlineAnalysis PCA and Independent Component Analysis ICA After newlinedimensionally reduced EEG signals are further processed by four different newlinepost processing techniques such as Hidden Markov Model newline newline