摘要
目的对于带约束的线性回归模型,y=Xβ+e,E(e)=0,cov(e)=σ2V,V>0,Rβ=0,给出了回归系数的有偏估计βR*(K)=(KM+I)-1βR*其中K=diag(k1,k2,…,kp)。方法应用比较的方法。结果与结论求出了在均方误差意义下βR*(K)优于β*R的条件,并讨论了其可容许性。
Aim To give a new biased estimation β*R(K)=(KM+I)-1β*R,where K=diag(k1,k2,…,kp) to restricted liner regression model y=Xβ+e,E(e)=0,cov(e)=σ2V,V0,Rβ=0.Methods Comparison method is used.Results and Conclusion The conditions in which β*R(K)=(KM+I)-1β*R and β*R(K) is superior to β*R in terms of mean square error are obtained,and their admissible estimations are discussed as well.
出处
《宝鸡文理学院学报(自然科学版)》
CAS
2011年第2期17-18,共2页
Journal of Baoji University of Arts and Sciences(Natural Science Edition)
关键词
有偏估计
广义最小二乘估计
可容许估计
biased estimation
generalized least squares estimation: admissible estimation