A Novel Approach for Classification of EEG subjects using Hybrid Machine Learning algorithm

Abstract

Recurrent Epileptic seizures are neurological events described by irregular electrical activity in brain, detectable through electroencephalogram (EEG) signals. Accurate detection of seizures is crucial for timely medical intervention by effective management and treatment. This paper presents end-to-end approaches to detect seizures from EEG subjects using K-Nearest Neighbour (KNN), Support Vector Machine (SVM), Decision Tree (DT) and a proposed Hybrid ML model. BONN Dataset is used to collect EEG subjects in this study. Our developed KNN, SVM and DT algorithm attained an accuracy of $\mathbf{9 4 \%} \% \mathbf{9 6. 6 7 \%}$ and $\mathbf{9 6. 6 6 \%}$. The developed hybrid algorithm improved the performance and accomplished an accuracy of $\mathbf{9 7. 3 3 \%}$. The developed end-to end models are very robust, effective and helps in automated seizure detection.