Gauss-Markov条件下最小二乘估计的强相合性

设Y_i=x＇_iβ+e_i,1≤i≤n为线性模型,β_n=(β_n1,…,β_np)＇为β=(β₁,…,β_p)＇的最小二乘估计,以u_n记(sum from i=1 to n(x_ix＇_i))的(1,1)元,v_n=u_n^-1.证明了在Ee_i=O且{e_i}满足Gauss-Markov条件时,v_i→∞及sum from i=2 to ∞(v_i^-2(v_i-v_i-1)log~2i<∞)为β_n1强相合的充分条件,且对任何ε_n→0,v_i→∞及sum from i=2 to ∞(ε_iv_i^-2(v_i-v_i-1)log²i<∞)已不再充分.提出了β_n1强相合的一个充要条件,它把β_n1强相合归结为正交随机变量级数的收敛问题.     

半参数回归模型小波估计的强逼近 *

考虑半参数回归模型y_i=x^T_iβ +g( t_i ) +e_i,i=1 ,2 ,… ,n ,其中 β∈R^d 为未知回归参数 ,g(·)为 [0 ,1 ]上的未知Borel函数 ,{x^T_i}为R^d 上的随机设计 ,{t_i}为常数序列 ,{e_i}为i.i.d .随机误差 ,Eei=0 .在适当的条件下 ,证明了 β和g(·)的小波估计 ^{^}β和^{^}g(·)的强相合性 ,并且得到了^{^}β和^{^}g(·)的强相合速度 .     

关于线性模型中误差方差估计的收敛速度的若干注记

For linear models y_j=x′_jβ+e_j(j=1,…,n,…) satisfying e₁,e₂,…i.i.d.     

线性模型中误差方差估计的指数收敛速度

Suppose given a linear model y_j=x'_jβ+μ_j, j=1,2,…, The random errors all have a mean zero and unknown variance σ², 0<σ²<∞. Let σ_n² be the estimate of σ² based on the residual sum of squares and calculated from (x_j, y_j), j=1,…,n. In this paper we show that if μ₁,μ₂,…, are independent but not necessarily identically distributed, and some further conditions on {μ_j} and (x₁|…|x_n) are satisfied, then for any ε>0 there exist constant ρ_ε, 0<ρ_ε<1, Such that P(|σ_n²-σ²|≥ε)=O(ρ_εⁿ).     

关于线性模型中误差方差估计中心极限定理余项的非一致性估计

in linear models, the error varianceσ2 of random errors is usually estimat-ted by the residual sum of squares (divided by a su ita ble degree of free-dom), based on the first n observation, (denote it by σ_n²). it is well known t that under certain conditions, the distribution of this estimate, when standardised, converges to the standard normal distribution. In this paper, it is shown that |G_n(x)-Ф(x)|=O(n^-δ/2(1+|x|)^-(2+δ)). when the errors are indepedent (maynot be identically distributed) and their 4 + 2δ order moments exist, where G_n(x) is the distribution of (σ_n²-σ²/(varσ_n²)^1/2,Ф(x)=1/(2π)^1/2∫_-∞^xe^-r2/2dt.     

类2的常数联络黎曼空间

A non-flat Riemannian space V_n is called Riemannian space with constant connection if its Christoffel symbols of the second kind are constant in some coordinate system {xⁱ}. Following G. Vranceanu [3], a Riemannian space V_n with constant connection is said to be of genus p, if the components of the fundamental tensor in the coordinate system {x^{i/sup>} can be written in the form g_ij=c_ijat_a(a=1,…,p) where c_ija are constant, and p is least.In this paper we prove the following}     

非参数回归函数最近邻估计Lp收敛的速度

Let (X,Y) be a R^d×R¹-valued random vector with E(|Y|)<∞,m(x)=E(Y|X=x) be the regression funvion of Y with respect to X.Suppose that (X_i, Y_i),i=1, …,n, are iid samples drawn from (X,Y). It is desired to estimate m(x) based on these samples,everoye discussed in 1981 (see [2]) the pointwise L_p-convergence of the nearest neigthoor estimate m_n(x) (see (5) of the present paper). In this article we further study the rate of this convergence.It is shown that if there exists p≥2 such that E |Y|^p<∞,then E|m_n(x)-m(x)|^p=O(n^-p/(d+2))a.s. for suitabte choice of the weighte C_m (see(4) of the present paper).     

概率密度的均匀核估计与近邻估计的均方误差

In this paper we consider the Mean Square Error (MSE) of two uaual estimates of density function f(x) at a point x: The uniform kernel estimate f_n(x) and the NN estimate fn(x). we- show that when f is differentiable for sufficiently high order at x. these MSE can be expanded in a form E(f_n(x)-f(x))²=A₁(x)n^-4/5 +A₂(x)n^-1+A₃(x)n^-6/5+…;E(f_n(x)-f(x))²=B₁(x)n^-4/5 +B₂(x)n^-1+B₃(x)n^-6/5+… And if we suitably choose the parameters in f_n and f_n to make A₁(x) and B₁(x)to assume its minimunm value, then we also, have A₂(x) =B₂(x) but A₃(X) differs form B₃(X). This result shows that while the two estimates are not identical with respect to MSE. each one can be superior to the other in various special cases.     

部分线性回归模型中的自适应估计

考虑半参数回归模型y_i=x_iβ+g(t_i)+e_i,1≤i≤n;其中g(·)是未知函数,β是待估参数,e_i是随机误差. 本文首先在(x_i,t_i)是i.i.d设计点列下,当g(·)的估计取一类核估计序列时,构造了β的自适应估计β_n,同时给出了β_n及g_n^*(g的最终估计)的渐近最优强弱收敛速度;然后在(x_i,t_i)是固定非随机设计点列下,当g(·)的估计取一般非参数估计(包括常见核估计和近邻估计),讨论了β的最小二乘估计的渐近正态性.     

一类随机场越界概率的大偏差

For Y₁, Y₂,…i .i .d . with Y₁～N(μ, 1) and S_n=(?)Y_i, the large deviations are obtained for theprobabilities that(?) conditionally given (i) S_m=0, and (ii) S_mi=ξ. Applied these results to the double change points model with some nuisance parameters, we developed the large deviation for the significance level of the likelihood ratio test.