半参数模型中参数的最优线性无偏差分估计

Optimal linear unbiased differential estimation of parameters in semiparametric model

考虑半参数回归模型y=Xβ+g(t)+ε,在此模型中引入差分算子,得到差分半参数模型,并根据风险函数算出其中参数的最优线性无偏差分估计。同时给出差分模型下参数的最优线性无偏估计的相对效率及其上下界。最后通过模拟算例来说明理论结果优越性。

In this paper,a semiparametric regression model is considered.In this model,a difference operator is introduced to obtain a differential semiparametric model,and the optimal linear unbiased differential estimate of the parameters is calculated according to the risk function.The relative efficiency of the optimal linear unbiased estimation of the parameters under the difference model and its upper and lower bounds are given.Finally,the superiority of the theoretical results is illustrated by simulation example.