Abstract—One of the most severe neurological conditions that abruptly changes a person’s way of life is epileptic seizures. Recent diagnostic approaches have concentrated on creating Electroencephalogram (EEG) methods based on machine/deep learning model, with the goal of creating new and efficient technologies for managing epileptic seizures. It is a challenging task to identify the seizure and seizure-free states of an epileptic patient by classifying EEG signals into ictal and interictal classes. Many machines learning-based approaches to analyzing and interpreting EEG signals for the aim of accurate categorization were previously introduced. Still, it is challenging to obtain comprehensive information on these dynamic biological signals, nevertheless, due to the non-linear and non-stationary nature of EEG data. This paper aims to develop an automated epileptic seizure diagnosis system with the use of advanced feature extraction and classification techniques. Here, the Maximum Overlap Discrete Transform (MODT) approach is used to extract the epileptic seizure-related features that are most valuable. The Redone Butterfly Optimization (RBO) technique is used to reduce the dimensionality of features in order to increase classification accuracy. The Gaussian Kernel Radial Network (GKRN) is used to precisely forecast the seizure and classify its proper class. To compare and validate the outcomes of the MODT-GKRN framework, a variety of measures and benchmark datasets have been used in this study.
 
Keywords—Electroencephalogram (EEG), machine learning, Maximum Overlap Discrete Transform (MODT), Redone Butterfly Optimization (RBO), Gaussian Kernel Radial Network (GKRN)


Copyright © 2023 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.