文献类型：期刊文献

英文题名：Performance evaluation for epileptic electroencephalogram (EEG) detection by using Neyman-Pearson criteria and a support vector machine

作者：Wang Chun-mei[1];Zhang Chong-ming[1];Zou Jun-zhong[2];Zhang Jian[2]

机构：[1]Shanghai Normal Univ, Dept Elect Engn, Shanghai 200234, Peoples R China;[2]E China Univ Sci & Technol, Dept Automat, Shanghai 200237, Peoples R China

年份：2012

卷号：391

期号：4

起止页码：1602

外文期刊名：PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS

收录：;EI(收录号：20120114652967);WOS:【SCI-EXPANDED(收录号:WOS:000300459700066)】；

语种：英文

外文关键词：EEG; Epileptic EEG; Discrete wavelet transform; Approximate entropy; Support vector machine (SVM); Neyman-Pearson criteria

摘要：The diagnosis of several neurological disorders is based on the detection of typical pathological patterns in electroencephalograms (EEGs). This is a time-consuming task requiring significant training and experience. A lot of effort has been devoted to developing automatic detection techniques which might help not only in accelerating this process but also in avoiding the disagreement among readers of the same record. In this work, Neyman-Pearson criteria and a support vector machine (SVM) are applied for detecting an epileptic EEG. Decision making is performed in two stages: feature extraction by computing the wavelet coefficients and the approximate entropy (ApEn) and detection by using Neyman-Pearson criteria and an SVM. Then the detection performance of the proposed method is evaluated. Simulation results demonstrate that the wavelet coefficients and the ApEn are features that represent the EEG signals well. By comparison with Neyman-Pearson criteria, an SVM applied on these features achieved higher detection accuracies. (C) 2011 Elsevier B.V. All rights reserved.