基于小波变换平滑主成分分析

小波变换具有很强的信号分离能力，很容易把随机噪音从信号中分离出来，从而提高信号的信噪比。本文把小波变换引入到因子分析中，提出了基于小波变换平滑主成分分析，该算法既保留普通主成分分析的正交分解，又具备了小波变换的信号分离能力。模拟数据和实验数据的结果表明，该算法具有从低信噪比的数据中提取出有用信息，并提高信号的信噪比。迭代目标变换因子分析处理实验数据的结果表明，基于小波变换平滑主成分分析的处理结果优

Wavelet Transfer Based Smoothing Principal Component Ana lysis

Wavelet transfer (WT) is a powerful technique in signal separation. It is very easy for WT to separate random noise from useful signals, and to increase signal to noise ratio. By introducing the WT into factor analysis (FA) technique, a novel algorithm named wavelet transfer based smoothing principal component analysis (WTBSPCA) was proposed. The algorithm involves both the orthogonal decomposition of common principal component analysis (PCA) and the smoothing ability of WT. A simulated data set and an experimental data set were investigated by this method, the results showed that WTBSPCA could extract useful signals from the data, with low signal to noise ratio, and enhance the signal to noise ratio. The results of iteration target transfer factor analysis (ITTFA) demonstrated that the WTBSPCA was superior to PCA.