Epileptic Seizure Classification Using Adaptive Sine Cosine Algorithm-Whale Optimization Algorithm Optimized Learning Machine Model

Epileptic seizure leads to the unconsciousness of the brain due to the lack of sleep, toxic consumption mainly. Now a days the death rate becomes high due to the negligence of the people who suffered from the seizure. The diagnosis of epileptic seizure at the early stage is essential. The manual diagnosis of detection and classification of seizure is difficult for radiologists. Several researchers have proposed automatic detection and classification of seizure, but somehow failed in detecting and classifying seizures related the computational time and accuracy. We are proposing a novel hybrid using Adaptive Sine cosine Algorithm-Whale Optimization Algorithm optimized Extreme Learning Machine (ASCA-WOA-ELM) model for classification of epileptic seizure. The hybrid ASCA-WOA technique is proposed to optimize the weights of the ELM model to improve the performance of the conventional ELM model. The EEG signals from University of Bonn dataset are considered for the research. First, the statistical features are extracted from the EEG signals using wavelet transform. The ASCA-WOA-ELM is fed with features for classification. The proposed ASCA-WOA method's uniqueness is shown by optimizing benchmark functions. The performance measure parameters such sensitivity, specificity and accuracy are evaluated from the proposed ASCA-WOA-ELM model. The ASCA-WOA-ELM model achieved 99.42% accuracy, 99.47% specificity, and 99.53% sensitivity. Further, the computational time of 21.2841 seconds achieved by the proposed ASCA-WOA-ELM model. The comparison results with other optimized models such as SCA-ELM, WOA-ELM, ASCA-ELM, WOA-ELM, along with the proposed ASCA-WOA-ELM model are presented