Prediction of Epileptic Seizures Through Machine Learning Classification Methods

This work aims to improve the identification and classification of seizures and epilepsy by means of electroencephalogram (EEG) data processing, therefore addressing the significant demand for accurate and efficient diagnostic tools in epilepsy control. We investigate six machine learning models-Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), k-Nearest Neighbours (k-NN), Logistic Regression (LR), and Naïve Bayes (NB)-to identify the most effective technique for epileptic seizure prediction. A powerful preprocessing pipeline was employed to lower noise while preserving basic EEG signal features, hence guaranteeing best input quality for model training and evaluation. Key criteria including accuracy, precision, recall, F1-score, and Area Under the ROC Curve (AUC) were used in evaluations of models. DT and RF regularly beat other models, with exceptional accuracy, balanced recall and precision, outstanding F1-scores, and good AUC values. These results show their suit for applications including seizure prediction and monitoring that demand a trade-off between false positives and false negatives. The study also underlines the challenges of preprocessing and the need of thorough noise reduction to retain significant information.