摘要
考虑回归模型y_i=t_iβ+g(x_i)+e_i,i=1,2,…,n,其中g(·)是定义在R′=(-∞,+∞)上的未知函数,e_i是未知干扰,基于g(·)的估计取一类核和近邻估计,证得了β和g(·)的估计最优收敛速度。
Consider the regression model y_i=t_iβ+g(x_i)+e_i for i=1, 2,…,n. Here g is an unknown function in R'=(—∞,∞), e_i is an unobserved disturbance. Based on the kernel and nearest neighbor estimators of g(·), some optimal convergence rates about the estimators of β and g(·) are obtained.
基金
This project supported by NNSFC under the contract 18901001
关键词
半参数
回归模型
最小二乘估计
semiparametric regression model, least-square estimate, optimal convergence rate
