具有不同光滑变量的变系数模型

Varying-coefficient Regression Models with Different Smoothing Variables

本文探讨具有不同光滑变量的变系数模型的建模、估计和估计的渐近性．首先,从实际出发建立模型;然后,使用局部线性方法给出模型中未知函数的初始估计,再使用平均方法,给出它们的平均估计;进—步,给出这些平均估计的渐近正态性．两个模拟例子说明这一估计方法是有效的．

In the paper, the varying-coefficient regression models with different smoothing variables in different coefficient functions are discussed. The averaged estimates of the coefficient functions are given by averaging the sample on the initial value obtained by local linear technique. Their asymptotical normalities are studied. Two simulated examples show that the procedure is effective.