线性等式约束的奇异线性模型的Liu型估计

Liu-type Estimator on Singular Linear Model with Linear Equation Restriction

考虑带等式约束的奇异线性模型的参数估计,为了克服复共线性问题,提出一个新的Liu型估计;同时给出这个估计的一些性质,并且得到了这个新的Liu型估计在均方误差矩阵准则(MSEM)下优于约束最小二乘估计的充要条件;得到在均方误差(MSE)准则下新估计优于约束最小二乘估计的充分条件.

This paper considers parameter estimator on singular linear model with equation restriction, presents a new Liu-type Estimator in order to overcome collinearity, gives some properties for this Estimator, obtains the sufficient and necessary condition for this new Liu-type Estimator superior to restricted least square estimator under mean square error matrix criterion and also receives sufficient condition for the new Estimator superior to restricted least square estimator in terms of mean square error criterion.