强混合样本回归函数估计的强相合性

STRONG CONSISTENCY OF ESTIMATES OF REGRESSION FUNCTION FOR STRONGLY MIXING SAMPLE

本文基于强混合样本，给出回归函数核估计的强相合性，全面地改进了胡舒合（１９９５）所得的相应的初步结果．并建立了与独立样本相当的最优收敛速度，所得结果也包括随机窗宽和密度函数核估计．

In this paper, under strongly mixing sample, we derive the strong convergence rate of deviation of the estimation from the true regression function. The best results corresponding to that for an i.i.d sample are obtained.