Epileptic Seizure Detection Using Feature Importance and ML Classifiers

Abstract

Epilepsy seizure (ES) monitoring and detection are only two examples of the many problems that maybe addressed by combining the Internet of Medical Things (IoMT) with machine learning (ML)techniques and cloud computing services. Epilepsy, a potentially fatal neurological disorder, is aworldwide problem that poses a significant threat to human health. There is an urgent need for areliable way of identifying epileptic seizures in their early stages to save thousands of epileptic patientsevery year. With the use of IoMT, several medical treatments, such as epileptic monitoring, diagnosis,and other procedures, may be performed remotely, hence lowering healthcare costs and enhancingservice quality. EEG datasets have made use of feature importance-based data reduction to address theproblem of a high number of data points and improve the delivery of service to the end user. In thisarticle, we use the feature importance method by applying two popular machine learning techniquesextra tree classifier (ETC) and the extreme gradient boosting classifier (XGBoost). Finally, theperformance of a number of tests is evaluated using experimental data from Bonn University. Alsoachieved is a comparison of the two approaches used. The collected findings demonstrate the efficacyof the XGBoost technique and its greater accuracy in comparison to the ETC strategy