Classification of EEG Signals using WPT, MGWO and Rule Based Classifiers

Abstract

The essential intent of the purported work is to develop an accurate automated seizure detection model for the performance evaluation of epileptic patients in an improved manner. Long data sets of EEG signals are recorded for a long duration of time which has taken from PhysioNet CHB-MIT EEG dataset for this experimental work. Six types of elements are excerpted from EEG signals by using WPT method. By using this feature extraction method, variance of monotonic amplitude, Mean of joint instantaneous amplitude and mean monotonic absolute amplitude as features are extracted . These features are inputted to each of the six classifiers for validation of the proposed method. Here, Modified Grey Wolf Optimization technique is used to optimize the parameters of the classifiers. Then, all the features are combinely inputted to the rule based six number of classifiers to detect normal and seizure EEG segments. The developed seizure detection WPT- Naive-Bayes method achieved excellent performance with the average Accuracy, specificity, sensitivity, G-mean, positive predictive value, and Mathews correlation coefficients as 97.24%, 97.34%, 97.13%, 98.1%, 96.99%, 97.66% respectively The average area under curve (AUC) is approximately 1. The proposed method is able to enhance the seizure detection outcomes for proper clinical diagnosis in medical applications.