基于离散小波变换的脑电信号睡眠分期研究

Research on sleep staging based on discrete wavelet transform for electroencephalogram

提出基于脑电信号(EEG)的睡眠分期研究。利用离散小波变换(DWT)的db8小波分解得到的细节分量作为信号新的表达,把各个细节分量能量作为特征,建立带高斯径向基核函数(RBF)的非线性支持向量机(SVM)模型。研究发现,其对睡眠分期研究的方案是可行的,满足模型对泛化能力的要求。

This paper proposes a research for sleep staging based on electroencephalogram （EEG）. By using the discrete wavelet transform （DWT）, the detail components by db8 wavelet decomposition are new expression of signal, whose energies are the feature. It establishes a nonlinear support vector machine（SVM） model with Gaussian radial basis kernel function（RBF）. The study shows that the program of this paper is practicable for sleep stage research, which meets the requirements for model generalization ability.