变量线性相关对主成分方差贡献率的影响

Impact on the Variance Contributes of the Principal Component Made by the Variable Linear Dependant

研究变量组个数、变量线性相关、线性无关和变量方差的改变,对矩阵特征值、主成分方差贡献率的影响.理论分析证明,线性相关变量个数的增加和变量方差的增大,都能使特征值增大,第一主成分方差贡献率增大,其它主成分方差贡献率减小.

This thesis studies the impacts,made by the number of the Variant set,the Variable Linear dependent,Linear independent and the change of the Variable average square.On the eigenvalues of matrix and the Variance contributes of the principal component,the analyses have proved that the increase of the number of the linear dependant variable and the variable average square increases both the eigenvalue and the variance contributes of the first principal component.However,it decreases the contributes of the other principal component.