Improving Seizure Detection by Integrating Feature Selection and Classification Techniques

Understanding the diverse triggers of epileptic seizures, which encompass head traumas, strokes, brain tumors, genetic conditions, developmental issues, and metabolic abnormalities, is crucial for their effective management and prevention. Identifying and predicting seizures can go a long way in managing and preventing them, especially given that they affect millions of people each year and are a major public health concern among neurological illnesses. The implementation of time-frequency domain feature extraction algorithms for seizure identification utilizing Electroencephalogram (EEG) signals is the subject of this study. This study investigates the application of the Discrete Wavelet Transform (DWT) in combination with other features to capture the temporal and frequency dynamics inherent in EEG signals. Seizure patterns are classified by using various Machine Learning Classifiers This integrating method of feature extraction with classification resulted in an accuracy of 97.70%, proving its efficacy in identifying seizures. This study adds to the field of seizure detection by proposing a viable strategy for improving early seizure prediction and diagnosis.