本文分析了小离差主成分与线性回归的几何意义,研究了利用小离差主成分建立线性方程或方程组.解决经济计量模型中参数估计问题,并得到了比较满意的结果。这种方法还具有能确定多重共线性、计算相对拟合度、减少回归次数、综合利用样本资料信息等有点。
This paper analyses the geometric meaning of small variance principle composition and linear regression, researches how to solve parameter estimation problem in economic measuring models through establishing linear equation or equations by small variance principle composition, and obtains satisfactory results. The method has the following advantages: deciding muhi-collinearity, computing opposite degree of fitting, reducing frequency of regression, utilizing synthetically sample data etc.