An Automated EEG Signal Analysis Using Discrete Wavelet Transform and Advanced Machine Learning Models

Epilepsy, a complex disorder characterized by recurrent seizures and neurological manifestations, affects a significant global population. Diagnosis relies on brain imaging and electroencephalogram (EEG) assessments to comprehend seizure patterns and their severity across patients. This study introduces an automated approach using 54 discrete wavelet transform (DWT) for EEG signal analysis, comparing classical classifiers—random forest (RF), decision tree (DT), Naive Bayes (NB)—with modern deep learning models like CNN, LSTM, and dense ANN for epileptic activity detection. The versatile DWT, extracting both spectral and temporal information from EEG signals, shows promise in delineating epileptic patterns. Evaluation across models highlights the superior performance of CNN and LTSM, achieving exceptional accuracy rates of 0.99 and 0.98, respectively. Precision, recall, and F1-score metrics reinforce the effectiveness of these models in accurately detecting epileptic activity, signifying the potential of advanced techniques in EEG-based epilepsy diagnosis.