Epileptic Seizure Classification Using Wavelet -Based Features and Evolutionary Feature Selection with Machine Learning Classifiers

Epilepsy is a chronic neurological disorder characterized by recurrent, unpredictable seizures and is commonly diagnosed through electroencephalogram (EEG) signal analysis. Accurate, automated detection of epileptic patterns in EEG signals is essential for timely diagnosis and effective treatment. In this study, we propose a robust framework for epilepsy detection leveraging Discrete Wavelet Transform (DWT) for feature extraction and Genetic Algorithm (GA) for optimal feature selection. EEG signals are first preprocessed and decomposed into multiple sub-bands using DWT to extract temporal, spectral, and time-frequency domain features. These features are then optimized via GA, which reduces dimensionality by selecting the most discriminative attributes. The selected features are then fed to Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Artificial Neural Network (ANN) classifiers. The proposed method was evaluated using the Bonn EEG dataset and validated through ۱۰-fold cross-validation. Results demonstrate that the SVM classifier achieved the highest accuracy of ۹۷.۸% and an AUC of ۰.۹۸۹. Our approach significantly outperforms traditional methods, showing enhanced efficiency and reliability due to effective feature selection. This study underscores the potential of integrating wavelet-based feature extraction with evolutionary algorithms to develop intelligent EEG-based diagnostic systems for epilepsy detection.

Keywords:

Epilepsy detection , EEG , Discrete Wavelet Transform (DWT) , Genetic Algorithm (GA) , Feature Selection , SVM , Time-Frequency Analysis