Epilepsy Detection in EEG Signal using Recurrent Neural Network

Abstract

In this paper, we proposed a Recurrent Neural Network (RNN) for the classification of epileptic EEG signal. The EEG dataset is first preprocessed using Discrete Wavelet Transform (DWT) to remove noise and extract features. 20 eigenvalues features were extracted and used to train and test our model. Several experiments were conducted to obtain the optimal parameters for the model. Our model was then compared against Logistic Regression (LR), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Random Forest (RF) and Decision Tree (DT). From experimental results, the best generalization of 99% accuracy is obtained with RMSprop at 0.20 dropout and 4 hidden layers for our model. DT classifier performed second best with accuracy of 98% while RF performed the worst at 75% accuracy.