相依样本下变系数模型中函数系数导数的统计推断  

Statistical Inference for the Derivative of Coefficient Functions in Varying Coefficient Model with Dependent Sample

作　　者：韩玉[1] 王亮红[1] 

机构地区：[1]东北电力大学理学院,吉林吉林132012

出　　处：《东北电力大学学报》2013年第6期48-51,共4页Journal of Northeast Electric Power University

基　　金：东北电力大学博士科研启动基金(BSTXM-201216);吉林省自然科学基金(201115180;201215165)

摘　　要：考虑相依样本下变系数模型系数函数导数的估计问题。对变系数模型系数函数应用局部线性回归,利用加权最小二乘方法,给出系数函数导数加权最小二乘估计,并证明了所提出的估计量的渐近正态性。This paper considers the estimation of derivative of coefficient functions in varying coefficient model with dependent sample. By applying the local linear regression to coefficient functions of varying coefficient model,the weighted least square estimations of the derivative of coefficient functions are obtained and proved the asymptotic normal distribution of the mentioned statistics.

关 键 词：强混合序列 变系数模型 函数导数 渐近正态性 

分 类 号：O212.7[理学—概率论与数理统计]