Seizure Prediction using Machine Learning on EEG Signals

Introduction

This is project is an attempt to predict epileptic seizure onset using machine learning techniques on EEG signals.

Outline

1. Dataset

The CHB-MIT Scalp EEG Database has been used in this project.

2. Feature extraction

Features have been extracted from the dataset using mne and pyeeg

The extracted features are:

• Mean Variance
• Mean Kurtosis
• Mean Skewness
• Petrosian Fractal Dimension
• Hjorth Mobility
• Hjorth Complexity
• Mean Spectral Entropy

3. Classifiers

• SVM
• RNN
• R-LSTM