固定设计下半参数回归模型核估计的渐近性质

Asymptotic Property of the Kernel Estimator for Fixed desgin Semiparametric Regression Model

研究一类新的半参数回归模型回归函数的核估计问题,其中误差项为一阶非参数自回归过程.通过重复利用Watson-Nadaraya核估计方法,构造了回归函数及误差回归函数的估计量分别为β,g(·)和ρ(·),在适当的条件下,证明了估计量β,g(·)和ρ(·)的渐近正态性.

Under fixed-design, considering semiparametric regression model with first non- parametric autoregressive errors, we construct the N-W kernel estimators of parameter and nonparametric function β, g （.） and p （.）, under suitable conditions, we prove the strong con- srstency of β, g（.）and p （.）, and give the asymptotic normality.