ABSTRACT
Epilepsy is a genuinely obstinate neurological problem and can be detected by dissecting the brain signals created by neurons. Neurons are associated with one another in a mind-boggling approach to speak with human organs and develop signals. In this chapter, we have used a proficient extreme learning machine classifier, a feed-forward neural network extricating highlights. Highlights utilized the PyEEG Library for epileptic seizure recognition in electroencephalogram (EEG) signals. We examined the presentation of highlights exclusively and aggregately in changing size investigation windows, and a dataset was taken from the Bonn University—EEG dataset. The component extraction measure was applied to the dataset to build the grouping execution. Techniques like ELM neural network, K-nearest neighbors, support vector machines (SVM), Gaussian SVM, random forest, and Sigmoid SVM were utilized to arrange the extricated features. The best characterization exactness was obtained as 99.93%, with a sensitivity of 91.23%. Specificity was 94.3% when two classes were made for EEG classification seizure and non-seizure.
