摘要
考虑线性回归模型Y=Xβ+e,(1)Y和e都是n维随机矢量,且Ee=0,Eee^(τ)=σ^(2)I;X是(n×p)阶系数矩阵,β是p维参数矢量。依据Y对β所做的最小二乘估计为■(2)
In ridge regression analysis the most important thing is to find the minimum value of the variance function(E‖β(k)-β‖^(2)).But first of all,we must find out whether it is true that the variance function has one,and only one,minimum value.This note points to the opposite and shows that there are a few variance functions that have not only one but at least two minimum values.So this note is worth consideration in both the application and the research of ridge regression analysis.
出处
《科学通报》
1979年第14期625-627,共3页
Chinese Science Bulletin
