相关系数矩阵与多元线性相关分析

Analysis on Correlation Coefficients Matrix and Multivariable Linear Correlation

相关系数指度量两个随机变量间线性关系的无量纲指标,在研究了相关系数矩阵性质及其与多元随机变量线性相关性之间关系的基础上,提出多元线性相关系数的定义,用于衡量多个变量间线性相关强弱的无量纲指标.分析表明,所提多元线性相关系数能够较全面地反映变量间的线性相关强度.

Correlation coefficient is a non-dimensional index for measuring the linearity between two variables.On the basis of studying correlation coefficients matrix and its relation to the linearity among multiple variables,this paper proposed a definition of multivariable correlation coefficient,which is a non-dimensional index for evaluating the linear intensity among multiple variables.Analysis shows the multivariable correlation coefficient presented in the paper can comprehensively express the linearity among multiple variables.