半参数回归模型随机加权估计的渐近展开  

Edgeworth expansion of random weighting estimation in semi-parametric regression model

作　　者：任争争[1] 李佼瑞[1] 徐伟[1] 

机构地区：[1]西北工业大学应用数学系

出　　处：《兰州理工大学学报》2005年第4期120-123,共4页Journal of Lanzhou University of Technology

基　　金：国家自然科学基金(10333020)

摘　　要：指出对简单的半参数回归模型y=xβ+g(t)+ε,在进行参数估计时,为避免利用Bootstrap法需进行重新抽样,应用随机加权法构造了涉及参数β的最小二乘随机加权估计量n,并且在适当的条件下获得了n分布的Edgeworth展开,其收敛速度可达到O(n-(1)/(2)).In order to avoid the compulsory resampling with Bootstrap method when the parameter estimation is performed with a simple semi-parametric regression model y=xβ＋g（t）＋ε, a least square randomweighting estimation quantity λn concerned with a parameter β was constructed by using random weightingmethod,and,in an adequate condition, the Edgeworth expansion of λn distribution was obtained with convergence speed of the order of O（n^-1/2）.

关 键 词：半参数回归模型 随机加权法 EDGEWORTH展开 收敛速度 

分 类 号：O212.1[理学—概率论与数理统计]