COMPARISON OF VARIOUS CLASSIFIERS IN EARLY DETECTION OF SEIZURES USING DEEP LEARNING

Abstract

The epilepsy is a neurological disorder that occurs due to the sudden recurrent episodes of sensory disturbance in the brain. It results in the loss of consciousness or convulsions that is merely associated with abnormal electrical activity occurring in the brain. That is mainly due to the disorder in the functionality of brain which adversely impacts the health of the patient. The standardized tool that helps in the detection of the brain activities and measuring the signals for epilepsy seizures is Electroencephalograms (EEG). The EEG signals are extensively used in the detection of seizures. The two techniques used for the detection of seizures in the EEG Signals are machine learning and deep learning. Two models have been broadly compared by performing the real time analysis on the EEG Signals by using both the methods and the results have been discussed in the research paper. The operations that have been from extracting the EEG Signals in the time, frequency domain. Alongside this, the two basic problems with the signals have been tried to solve and reduce to a larger extent those are noise removal and feature extraction.  The accuracy and precision of utilizing both the techniques have been measured and accordingly the results have been obtained. The efficiency achieved by using the methods out to be 94.6% which is highly effective in predicting the seizures.