【摘要】 局部线性嵌入算法将高维信息通过变换降到低维数的特征空间中,从而压缩数据突出信号主要特征。该算法很好地弥补了线性降维不能发现数据集非线性结构的不足。本文详细介绍了局部线性嵌入算法的基本原理和运算步骤,并将该方法应用于混有高斯白噪声的ECG信号降噪和混有弱冲击正弦信号的特征提取中。处理结果表明,局部线性嵌入算法不仅可以处理线性信号,还能较好地处理非线性信号,具有较好地工程推广价值。更多还原

【Abstract】 High dimension data can be projected to low dimension space by using locally linear embedding algorithm (LLE). Hence, the compressing data can enhance the main feature of signal. LLE is able to find nonlinear structure of data set. This paper describes the basic principle and realization steps of LLE in detail. Then this algorithm is adopted to reduce the noise of ECG and extract the shock feature of simulated signal. The result shows LLE can process the linear signal as well as the nonlinear signal. Consequently,LLE is an effective method to apply in signal processing.更多还原