Using FFNN Classifier with HOS-WPD Method for Epileptic Seizure Detection

Abstract

Epilepsy is a damage in brain cells that cause a disturbance in the electrical signals of the brain, leading to nervous system disorder. Studies show that about 1% of the world's population suffers from this disorder [1]. Epilepsy can be diagnosed by studying the electroencephalogram (EEG)signals,, i.e. the electrical signals emitted from the brain and represent its activity. This paper proposed an epileptic seizures detection based on analysis of EEG signals. The detection is carried out firstly recording the EEG signals using the EEG device. The noise is then eliminated before features extraction process is carried out using HOS-WPD. These features are then used to train two classifiers, namely Navie Bayes and FFNN, by which the signals are classified as either benign or seizure. Experimental evaluation was carried out to compare the detection performance of both algorithms in terms of Precision, Recall, and Accuracy and using MIT BIH Dataset.