The main unrelated signals can be obtained by reducing its dimensions through singular value decomposition for the covariance matrix using the traditional principal component analysis, so that the purpose of reduction data, revealing the relationship between variables and interpretation of data can be realized. The variance of input signal affects the results of analysis, so the method also exposure its insufficiency. In the based of this reason, the improved method of the principal component analysis was be researched which based on correlation function.%传统的主成分分析方法是针对原始数据的协方差矩阵进行奇异值分解,将原来多个变量构建为少数几个互不相关的主成分的一种统计方法,最终达到数据化简、揭示变量间的关系和进行数据解释的目的.但是,对于传统的主成分分析方法,输入信号的方差对其分析结果影响较大.因此,本文开展以相关函数矩阵为基础的主成分分析方法研究.
展开▼
