EEG seizure classification based on exploiting phase space reconstruction and extreme learning

Epilepsy is the neural condition of the mind activity.Approximately 50 million persons currently live with epilepsy globally. Closely 80% of the Persons with Epilepsy (PWE) living in little- and middle-income countries. Analysis of epilepsy lies on the electroencephalogram (EEG) signals through the various behavior of brain disorders. Localizing the seizure points is generally is tedious and time consuming. Therefore one often to acquire and analyze ten or even hundred hours of recordings. Alternative method is to predict the seizure before they occur and allow the patients to take appropriate care. In this paper, EEG signal is decomposed into coefficients by Wavelet transform. Second, state space reconstrunction method is used to derive phase space trajectory by converting time space signal to phase space. Finally, the classification performance is associated by accrucay and training time with fuzzy system and extreme learning machine (ELM) in neural system. The comparative experimental results using publically available data and real time data are comprised to display the efficiency of the suggested work towards the examination of seizure gestures. The novelty of this paper includes the analysis, detection and discrimination of seizure events.