Utilizing Eeg Signals for Epilepsy Seizure Detection

Epilepsy is a central nervous system disease causing abnormal brain activity, odd behavior, and coma. It affects individuals of various ages and ethnicities, making seizure detection challenging due to differences in actions and brain signals. The rise in global infections has spurred the development of accurate techniques for early detection and improved treatment outcomes. The suggested method consists of three parts: preprocessing, feature extraction, and classification. Kid-specific EEG recordings from the CHB-MIT database are preprocessed to obtain a pure signal. Mel Frequency Cepstral Coefficients (MFCC) are then extracted as features. Categorization methods like Random Forest (RF) and K-Nearest Neighbor (KNN) are utilized for classification. Study findings demonstrate the potential of machine learning algorithms in enhancing epilepsy diagnosis using EEG data. KNN achieves up to 99.05% accuracy, while Random Forest achieves 95.02% accuracy in detecting epileptic activity. These results highlight the value of machine learning in epilepsy as diagnostic and therapeutic tools.