多变量数据分析及应用研究

在统计信号处理及其相关领域,多变量数据的描述和分析一直是人们广泛关注的研究课题.在现有的多变量数据分析方法中,基于二阶统计特性的主分量分析(PCA)和基于高阶统计特性的独立分量分析(ICA)是两种非常有代表性的方法.本文在简要介绍PCA和ICA基本原理的基础上,结合脑电消噪问题,对两种方法的性能和特点进行了较深入地比较.实验结果表明,在非高斯信号处理上,独立分量分析方法具有明显的优势.

Study of Multivariate Data Analysis and Its Application

In statistic signal processing and related areas, multivariate data analysis technique has attended wider attention in recent years. Principle Component Analysis(PCA) and Independent Component Analysis(ICA) are two representative of multivariate data analysis methods , The first is a decorrelation-based method and only uses 1st and 2nd order statistics of random variables; the second is a independence-based method and use higher-order statistics. The basic theories and algorithms are briefly introduced in this paper. The detailed comparison of PCA and ICA is studied with the application of noise removal of EEG. The experiment results show that ICA has obvious advantages in non-gauss signal processing.