Analysis of EEG Signal for Epilepsy Seizure Detection Using Machine Learning

Abstract

- The brain is a vital organ responsible for regulating various bodily functions. An electroencephalogram (EEG) is a powerful tool for
recording brain activity from the scalp. Analyzing EEG signals plays a key role in diagnosing and treating neurological conditions such as epilepsy, seizures, and other brain disorders. However, these signals are often affected by noise and artifacts, making analysis
challenging. Accurate evaluation of EEG signals is critical for reliable diagnosis. Numerous approaches have been developed to achieve high-precision analysis. In this study, Discrete Wavelet Transform (DWT) is utilized to pre-process EEG signals, breaking them down into five distinct frequency bands. Subsequently, features such as mean, standard deviation, root mean square (RMS), entropy, energy, and relative energy are extracted. These features are classified using machine learning algorithms, including Support Vector Machine (SVM), K- Nearest Neighbor (KNN), and Naive Bayes (NB). The findings reveal that SVM achieves the highest performance, with a classification accuracy of 99.5%, specificity of 100%, and sensitivity of 100% when distinguishing between normal and epileptic subjects