基于DD-DWT和Log-Logistic参数回归的癫痫脑电自动识别方法

发布时间：2018-02-26 08:27

本文关键词： 癫痫脑电双密度离散小波变换 Log-Logistic参数回归模型　出处：《仪器仪表学报》2017年06期 　论文类型：期刊论文

【摘要】：针对现有癫痫脑电(EEG)识别算法分类模式单一、普适性不强的问题,提出了一种新的基于双密度离散小波变换(DDDWT)和Log-Logistic参数回归(LLPR)的脑电信号自动识别方法。不仅利用了DD-DWT算法的分解特性,还建立了脑电信号的LLPR模型,并将二者有机的结合,从而更好的发挥算法的优势。滤波后脑电信号由DD-DWT进行6层分解,提取各子频带系数的小波域能量波形,并结合LLPR模型计算尺度参数α和形状参数β以表征信号,将构成的特征向量送入遗传算法(GA)优化的支持向量机(SVM)得出识别结果,从而实现脑电信号的自动识别。所提方法在处理A\D\E与AB\CD\E两种多模式脑电分类问题时,识别率分别为98.90%和97.75%。实验结果表明,所提算法更符合实际应用需求,可以较好地解决多类脑电信号识别问题,具有良好的普适能力和分类性能。
[Abstract]:Aiming at the problem of single classification pattern and low universality of the existing algorithms of EEG-based epilepsy recognition, A new method of EEG automatic recognition based on dual density discrete wavelet transform (DWT) and Log-Logistic parameter regression is proposed. Not only the decomposition characteristic of DD-DWT algorithm is used, but also the LLPR model of EEG is established, and the two models are combined organically. After filtering, the EEG signal is decomposed into six layers by DD-DWT, and the wavelet domain energy waveform of each subband coefficient is extracted, and the scale parameter 伪 and shape parameter 尾 are calculated with LLPR model to represent the signal. The feature vector is fed into genetic algorithm (GA) and the optimized support vector machine (SVM) is used to obtain the recognition result, so as to realize the automatic recognition of EEG signal. When the proposed method is used to deal with two kinds of multi-mode EEG classification problems, A\ D\ E and AB\ CD\ E, two kinds of multi-mode EEG classification problems are obtained, such as A\ D\ E and AB\ CD\ E respectively. The recognition rates are 98.90% and 97.75, respectively. The experimental results show that the proposed algorithm is more suitable for practical applications and can solve the problem of multi-class EEG signal recognition. The proposed algorithm has good adaptability and classification performance.
【作者单位】：吉林大学通信工程学院;
【基金】：吉林省科技发展计划自然基金(20150101191JC) 吉林大学研究生创新项目(2016092) 中央高校基本科研业务费专项资金(451170301193)资助
【分类号】：R742.1;TN911.23

【参考文献】