Epileptic seizures, caused by sudden and uncontrolled electrical activity in the brain, pose significant risks such as injuries and loss of physical control. Predicting seizures accurately can enable preventive measures and timely interventions, improving safety and quality of life for individuals with epilepsy. This study proposes a hybrid model combining neural networks and ensemble learning techniques to enhance seizure diagnosis using EEG signals. By integrating algorithms such as XGBoost (XGB), Support Vector Machines (SVM), Random Forest (RF), and Bidirectional Long Short-Term Memory (BiLSTM), the model achieves superior accuracy. Using the Bonn University open-access dataset, the proposed methodology attains an accuracy of 98.52%, precision of 97.37%, recall of 95.29%, and an F1-score of 96.32%. These results demonstrate the model's effectiveness in improving evaluation metrics and providing a robust framework for seizure prediction and diagnosis.

Keywords: Epileptic seizure diagnosis, ensemble learning, deep learning, EEG signals, XGB, SVM, RF, BiLSTM, seizure prediction.