φ混合误差下非线性模型LS估计的收敛速度

The Rate of Convergence of the Least Square Estimator in a Nonlinear Regression with ■-mixing Errors

本文对非线性模型Xn＝yn（θ）＋εn；获得了θ的LS估计的a．s收敛速度，推广和改进了PrakasaRao（［5］）的结果．

For nonlinear model Xn=gn(θ)+εn, we obtain the a.s. convergence rate of the LS estimatorθn, provided that {εn} is an φ-mixing errors with Eεi=θ, supiE|εi|t＜∝(for some t＞2),and other suitabe conditions, which generalized and improved Prakasa Rao'sresult (J. Mult. Anal. 14(1984),315-322).