AI-Powered Classification of EEG Signals for Seizure Detection in Neurological Disorders

The aim of this project is to explore the effectiveness of machine learning and deep learning techniques for classifying EEG signals in the context of neurological disorders, particularly focusing on seizure and non-seizure classification. The dataset used in this study consists of ECG signals, sampled at 178 time intervals, which are then transformed into key features like Power Spectral Density (PSD), entropy measures, statistical features, and peak frequency. Various machine learning models, including Random Forest (RF), LightGBM, Stacking Classifier, and deep learning models such as Convolutional Neural Networks (CNN) and LSTM combined with XGBoost, are applied to classify the signals as seizure or non-seizure. The model performance is evaluated based on classification accuracy and other metrics. The project also includes a web interface built with HTML, CSS, and Flask, allowing users to register, log in, and input their EEG signal data for prediction. The system provides predictions on whether a seizure is likely to occur, enhancing the potential for early diagnosis and intervention in patients with neurological disorders. This work demonstrates how machine learning and deep learning can be leveraged for the effective classification of EEG signals, contributing to advancements in healthcare.

Keywords:

EEG signal classification, neurological disorders, machine learning, deep learning, seizure prediction, feature extraction, Power Spectral Density, Random Forest, LightGBM, CNN, LSTM, XGBoost, Flask, healthcare, early diagnosis, classification models.