带ARMA误差的曲线回归模型的估计

An Estimation of Curve Regression Model with ARMA Series Error

考虑广义回归模型 yi=g( ti) +εi,1≤ i≤ n,其中 g( .)为 R上的未知函数 ,随机误差εi 是 ARMA( p,q)序列 .本文利用线性小波光滑的方法 ,讨论未知函数 g( .)的小波光滑及 ARMA( p,q)的参数估计 .

Consider the generalized curve regression model give by y i=g i(t)+ε i where {ε i} is ARMA series. Our goal is to estimate curve g(·) and estimation of AKMA series based on {y i} n i=1 , We apply wavelet smoothing to model and study the convergence of the wavelet estimation.