Quantitative EEG based on Renyi Entropy for Epileptic Classification

摘要

Analysis on Electroencephalogram (EEG)signal can provide important information related to theclinical pathology of epilepsy. Detecting the onset,prediction and type of seizures based on EEG signals isvery important to determine an appropriate treatmentfor the patients. However, EEGs have the highcomplexity with non-linear and non-stationarycharacteristics; hence, an analysis will be very difficultto do through a visual inspection. Signal processingapplications are, therefore, needed to make theinterpretation easier. In this study, we proposed amethod for EEG analysis based on signal complexity forthe epileptic EEG classification. The Renyi entropy wasused to extract the data of EEG features, which consistof seizure, interictal and normal features. Then, thesefeatures became the input to a classification algorithm.SVM (Support vector machine) classifier was applied todetermine the type of that epileptic EEG signal andachieved accuracy of 85 %. This study can be areference for neurology as an efficient method forepileptic EEG classification.