线性模型中M估计的渐近性质

线性模型中Ｍ估计的渐近性质吴耀华（中国科学技术大学数学系，合肥２３００２６）ＡＳＹＭＰＴＯＴＩＣＢＥＨＡＶＩＯＲＯＦＭ－ＥＳＴＩＭＡＴＯＲＳＩＮＬＩＮＥＡＲＭＯＤＥＬＳ￥ＷｕＹＡＯＨＵＡ（ＤｅｐａｒｔｍｅｎｔｏｆＭａｔｈｅｍａｔｉｃｓ，Ｕｎｉｖｅｒｓ...     

线性模型参数M估计的强相合性

本文研究线性模型中回归参数M估计的强相合性,给出一些较弱的充分条件.与相应结论比较,这里给出的条件对矩的要求有实质性的改进.     

线性模型参数M估计的强相合性

本文研究线性模型中回归参数　Ｍ　估计的强相合性，　给出一些较弱的充分条件．　与陈希孺和赵林城的专著［１］中相应的结论比较，　这里给出的条件对矩的要求有较大的实质性改进．     

线性模型参数M估计的强相合性

研究了线性模型中回归参数 M估计的强相合性 ,在 dn=o( 1 / logn)情形下 ,给出了较弱的充分条件 .与 [1 ]中相应结论比较 ,我们将有界性条件改善为矩母函数存在性条件 ,因此作了实质性的改进     

随机截断下部分线性模型中参数估计的渐近性质

考虑部分线性回归模型Yi＝xiβ g(ti) σiej,i＝1，2，…，n其中σi^2＝／f(ui)．当Yi因受某种随机干扰而被右截断时，就截断分布巳知的情形，利用所获得的截断观察数据构造了β，g，f的估计量β^～n，g^～n，f^～n，并在一定条件下，证明了β^～n的渐近正态性，同时得到了g^～n，f^～n的最优收敛速度。     

误差为NOD线性模型参数M估计的强相合性

研究了随机误差为NOD序列的线性模型中回归参数β_0的M估计,在较弱的矩条件下证明了回归参数M估计的强相合性,推广和改进了陈和赵(1996),杨(2002)和Wu(2006)等的结果,推广了肖(2007)等的结果.     

误差为NSD序列的广义线性模型的M估计的强相合性

研究了以NSD序列(negatively superadditive dependent)为误差的广义线性模型,得到了未知参数的M估计.在较弱的条件下,利用指数不等式、NSD序列加权和的强收敛性和Borel-Cantelli引理等证明了未知参数M估计的强相合性.此结果推广了独立误差和NSD误差的线性模型的相应结果.     

半参数方差误差模型的估计

本文综合近邻权函数法及最小二乘法，用两阶段最小二乘估计的方法得到了半参数EV模型中参数的估计量及其强相合性，渐近正态性。同时也得到了非参数函数的估计量及其强相合性，一致强相合性。     

误差为线性过程时回归模型的估计问题

对一类非线性回归模型及线性模型，在误差是一个弱平稳线性过程及适当的条件下，获得了估计量的r-阶平均相合性、完全相合性和渐近正态性。     

线性约束下部分线性模型中估计的渐近性质

考虑回归模型yi=x′iβ+ g(ti) + ei, 0 ≤i ≤nr=Rβ其中(xi,ti)是固定非随机设计点列,xi=(xi1,…,xip)′,β=(β1,…,βp)′(p 1) ,g是定义在[0 ,1]上的未知函数,β是未知待估参数,0≤ ti≤1i,ei 是i.i.d随机误差,且Eei=0 ,Ee2i=σ2 <∞.r是一个J维向量,R是一个J* p列满秩矩阵,基于g的估计取一个非参数权估计,本文讨论了在线性约束下β的最小二乘估计的相合性及渐近正态性.     

部分线性自回归模型中误差方差伪最小二乘估计的渐近正态性

Consider the model Y_t=βY_t-1 g(Y_(t-2)) ε_t for 3＜=t＜=T.Here g is an unknown function,βis an unknown parameter,ε_t are i.i.d,random errors with mean 0 and varianceσ~2 and the fourth momentα_4,andε_t are independent of Y_s for all t＞=3 and s=1,2. Pseudo-LS estimators■_T~2,■4T and■_T~2 ofσ~s,α_4 and Var(ε_3~2)are respectively constructed based on piecewise polynomial approximator of g.The weak consistency of■4T and■_T~2 are proved.The asymptotic normality of■_T~2 is given,i.e.T~(1/2)(■_T~2-σ~2)/■_T converges in distribution to N(0,1).The result can be used to establish large sample interval estimates ofσ~2 or to make large sample tests forσ~2.     

Parameter Estimation for Stochastic Parabolic Equations: Asymptotic Properties of a Two-Dimensional Projection-Based Estimator

A two-dimensional parameter is estimated from the observations of a random field defined on a compact manifold by a stochastic parabolic equation. Unlike the previous works on the subject, the equation is not necessarily diagonalizable, and no assumptions are made about the eigenfunctions of the operators in the equation. The estimator is based on certain finite-dimensional projections of the observed random field, and the asymptotic properties of the estimator are studied as the dimension of the projection is increased while the observation time is fixed. Simple conditions are found for the consistency and asymptotic normality of the estimator. An application to a problem in oceanography is discussed.     

纵向数据下混合效应EV模型中参数估计的渐近性质

考虑纵向数据下混合效应EV模型。对带有惩罚项的Profile广义最小二乘方法进行了修正。利用矩估计法和ML-based EM算法给出了固定效应,随机效应以及协方差阵的估计。在一般的条件下,给出了固定效应估计的强相合性和渐近正态性,并对所提出的各种估计进行了模拟研究。模拟效果不错。     

缺失数据下线性模型回归系数估计的大样本性质

在响应变量满足MAR缺失机制下，我们分别研究了基于观察到的完全样本数据对、基于固定补足后的“完全洋本”和基于分数线性回归填补后的“完全洋本”得到的回归系数的最小二乘估计的弱相合性、强相合性及渐近正态性，我们还通过数值模拟，比较了基于上述估计得到的β的置信区间的优劣。     

Truncated Estimator of Asymptotic Covariance Matrix in Partially Linear Models with Heteroscedastic Errors

A partially linear regression model with heteroscedastic and/or serially correlated errors is studied here. It is well known that in order to apply the semiparametric least squares estimation （SLSE） to make statistical inference a consistent estimator of the asymptotic covariance matrix is needed. The traditional residual-based estimator of the asymptotic covariance matrix is not consistent when the errors are heteroscedastic and/or serially correlated. In this paper we propose a new estimator by truncating, which is an extension of the procedure in White. This estimator is shown to be consistent when the truncating parameter converges to infinity with some rate.     

Asymptotic normality and strong consistency of maximum quasi-likelihood estimates in generalized linear models

In a generalized linear model with q x 1 responses, the bounded and fixed (or adaptive) p × q regressors Zi and the general link function, under the most general assumption on the minimum eigenvalue of ZiZ'i,the moment condition on responses as weak as possible and the other mild regular conditions, we prove that the maximum quasi-likelihood estimates for the regression parameter vector are asymptotically normal and strongly consistent.