Classifying Epileptic Seizure Events from EEG Signals by a Lightweight Machine Learning Model

Because of the substantial interest and necessity, there has been much recent work in epileptic seizure detection using machine learning models. U sing un-intrusive measurements of brain activity such as electroencephalograms (EEG) has allowed for large datasets to be constructed and used for computational intelligence to identify seizure events within EEG data. In this paper, we use a publicly available EEG dataset to develop a lightweight Machine-learning supervised model (simple Decision Tree) to classify seizure events from brain waves. The performance of this developed model was compared with a complex ML model (Support Vector Machine). The cross-validated Decision Tree model performed better for seizure event classification with an overall accuracy of 91.17%. This lightweight model will allow for developing mobile applications and user comfort.