Epileptic seizure is a neurological disorder, which affects the activity of brain through sensations, and unusual behavior. However, detecting the epileptic seizure using Electroencephalography (EEG) signal is a challenging task in the clinical sector. Various seizure classification methods are introduced to diagnose the seizure, but automatic seizure diagnosis is a challenging issue in the research community. Hence, an effective seizure diagnosis method named atom grey wolf optimization (AGWO)-based deep stacked auto encoder is proposed to perform automatic epileptic seizure diagnosis. The deep stacked auto encoder encoded the input vector through multiple layers to generate the hidden stage representation and the effective classification using the auto encoder is enabled through the optimal training of the classifier using the proposed Atom grey wolf optimizer (AGWO). The proposed AGWO is the integration of the Atom Search Optimization (ASO) and Grey Wolf Optimizer (GWO). It is worth interesting to note that the proposed AGWO-based deep stack auto encoder performed the accurate epileptic seizure classification using the EEG signal features, which forms the input to the classification phase. The proposed classification approach attained better performance with the metrics, like accuracy, sensitivity, and specificity with the values of 94.103%, 91.98%, and 97.344% through varying the training percentage.
Epileptic seizure, Deep stacked auto encoder, Atom Search Optimization (ASO), Grey Wolf Optimizer (GWO), Electroencephalography (EEG)