LNQD序列回归函数小波估计的渐近正态性

On the Asymptotic Normality for LNQD Sequences of Wavelet Regression Function Estimators

本文考虑非参数固定设计回归模型Yi=g(ti)+εi,1<i<n,其中{ti}是非随机固定设计点,g(t)是回归函数,{εi}为平稳LNQD序列随机误差.在适当的条件下,用异于文献[5]的估计方法,讨论了函数g(t)的小波估计量的渐近正态性,得到了与文献[5]相同的结论.

This paper considers the nonparametric fixed design regression model Yi =g（ti）＋εi,（l〈i〈n）, where {ti} is non-random design points, g（t） is regression function, and {εi} is a strictly stationary linearly negative quadrant dependent sequences. Under certain conditions, using a method different from paper [5], the asymptotic normality for the wavelet estimator of g（t） is studied and the same conclusion as in paper [5] is drawn.