首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于EEMD和LS-SVM的癫痫脑电信号识别
引用本文:李营,;吕兆承.基于EEMD和LS-SVM的癫痫脑电信号识别[J].淮南师范学院学报,2014(3):8-11.
作者姓名:李营  ;吕兆承
作者单位:[1]淮南师范学院电气信息工程学院; [2]淮南师范学院物理与电子信息系,安徽淮南232038
基金项目:淮南师范学院自科研究项目“基于脑一机接口的不同思维作业脑电信号的特征提取和分类”(2011LK93q)
摘    要:针对癫痫脑电(EEG)信号的非平稳性和非线性,提出一种基于集合经验模式分解(EEMD)提特征并利用最小二乘支持向量机(LS-SVM)的脑电信号分类方法。首先利用EEMD将EEG信号分成多个经验模式分量,得到各阶本征模式分量(IMF),然后提取有效特征,最后用LS-SVM对其进行分类,实验结果表明,该方法对癫痫发作间歇期和发作期EEG的提特征后分类识别正确率达到98%。

关 键 词:癫痫脑电信号  集合经验模式分解  最小二乘支持向量机

Recognition of epileptic EEG signals based on EEMD and LS-SVM
Institution:LI Ying;LV Zhaocheng;
Abstract:This paper proposes an ensemble empirical mode decomposition(EEMD) and least squares support vector machine(LS-SVM) classification method for non-stationary and no:n- linear epileptic EEG signal. Firstly, EEMD was used to decompose interictal and ictal EEGs into multiple empirical mode components. The EEG of intermittent period and ictal was decomposed into some IMFs(intrinsic mode functions). Secondly, approximate entropies and energy entropy of effective IMFS were extracted as features.Finally, the EEG was classified with LS-SVM. The experiment indicated that the classification accuracy of the proposed feature extraction method for interictal and ictal EEGs reached 98%.
Keywords:epileptic EEG signal  ensemble empirical mode decomposition(EEMD)  least squares-support vector machine(LS-SVM)
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号