Epileptic seizure detection of electroencephalogram based on weighted-permutation entropy

摘要

Weighted-permutation entropy (WPE) is modified from permutation entropy (PE), which recently has been proposed as a measurement for nonlinear time series. To explore the efficiency of this method and the features of seizure electroencephalogram (EEG) segments, we investigate the application of WPE in the complexity analysis for epileptic seizure detection based on EEG. It is found that the calculated values of WPE are decreased during seizure segments in the contrast with seizure-free. Moreover, WPE provides a highly distinguishable feature for classifier to obtain more accurate results of automatic classification compared with PE by support vector machine (SVM).