一类离散分布参数的经验Bayes估计的收敛速度

In this paper we consider a family of discrete distributions f_θ(x)dμ(x), and suppose that the Bayes estimate of φ(θ) with respect to the priori distribution H∈H has a form dH(x) =(?)a_k(x)f(x+k)/f(x). where f(x)=∫f_θdH(θ) . we construct asequence of empirical Bayes estimates and establish its rate of convergence, and prove that under suitable conditions this rate of convergence can arbitrarily close to 1. we also give a counter-example to the main Theorem 2.1 of [5], and then declare that the "Theorem" does not hold.     

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.     

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).     

关于有限域中原根分布的注记

Let f₁(x) , …,f_r(x) denote quadratfree polynomials over GF(q) that are relatively prime in pairs and of degree≥1. We have proved that if q≥(k-1)²(2^rω(q-1)-1)², then all the f_i(ξ) are primitive roots of GF(q) for some ξ∈GF(q), where k =degf₁+…+degf_r . With extending the results obtained by Wang Juping, Sun Qi and Han Wenbao respectively, we have also given a quantitative analysis of Carlitz's result.     

随机中止试验的线性模型中参数估计的几个性质

When sample size is a sequence of r.v., the parametric estimates in linear models are interesting both theoretically and practically. For linear models Y_i=x_iβ+e_i,i=1,2,…where {e_i} are p-dimensional i .i .d.r.v′.s with Ee₁=0. Var e₁=σ²,0<σ²<∞. we consider estimate σ²(v_n) of error variance σ² and least squares estimate β(v_n) of regression coefficient β, here {v_n} is a sequence of r.v. with positve integers larger than P. In this paper some properties of estimates are discussed . Those are unbiasedness, consistency and normality . For σ²(v_n), we also obtain some orders for convergence to normal distribution.     

解在QK 型及Dirichlet 型空间中的线性微分方程

对于单位圆盘上系数函数是解析函数的复微分方程
f⁽ⁿ⁾+A_n-1(z)f^(n-1)+…+A₁(z)f''+A₀(z)f=0,
给出了方程的系数函数和解函数之间的关系, 即当系数函数A_j 满足给定的条件时, 方程的所有解属于Q_K型空间和Dirichlet 型空间.     

多元Szász—Mirkjan算子的一致逼近

本文研究了多元Szása—Mirakjan算子在C_2B(T)中的逼近性质,利用K—泛函,建立了等价的逼近定理.主要结果如下 定理设f∈C_2B(T),0a) ;(ii)‖S_n,m(f)-f‖_∞ =0(n^-a);(iii)a)‖f(x+tφ(x),y)-2f(x,y)+f(x-tφ(x),y)‖_∞ =0(t<     

三角域上Bernstein多项式的Lipschitz常数

设T是平面上以T₁,T₂,T₃为顶点的三角形,f(p)为定义在T上的函数,称Bⁿ(f,P):=(?)f(i/n,j/n,k/n)B_i,j,kⁿ(P),为f的n次Bernstein多项式,这儿B_i,j,kⁿ(P):(n!)/(i!j!k!)uⁱv^jω^k是Bernstein基函数,(u,v,w)是P关于T的重心坐标。 B.M.Brown等人对单变量的Bernstein多项式证明了如果f∈Lip_Aλ,0<λ≤1,则对所有的n,都有B^α(f,x)∈Lip_Aλ。本文的目的是对定义在三角域T:{(x,y):x≥0,y≥0,x+y≤1}上的Bernstein多项式证明了类似的结果: 设f(P)∈Lip_Aλ,0<λ≤1,则对所有的n,Bⁿ(f,P)∈Lip_(2^1/2λA)λ,并且,在一定意义上,常数2^1/2λA是最好的。 上述结果对于任意的锐角或直角三角形T,也是成立的。 最后还指出,当T可为钝角三角形时,则不存在同一常数C,使对所有的n和任意三角形T,有Bⁿ(f,P)∈Lip_cλ。     

关于控制算子的若干注记

Let B(H) be the set of all bounded linear operators on a Hilbert space H. An operator T∈B(H) is called dominant if (T-λ)(T-λ)^*≤M_λ²(T-λ)^*(T-λ),?λ∈C.The numerical range of T is difined by W (T) = {(Tx, x): ‖x‖ = 1, x∈H}. In Section 1 some new characteristic of dominant operators are given. If C = AB - BA, we prove that O∈W(C)^- then A is a dominart or φ-quasihy ponor-mal. In Section 2 we prove that O∈σ_e(△_A^σ) if A is a dominant, where(?), we also prove that if A∈B(H) is a norm attaining Ф-quasihyponormal, then A has a non-trivial invariant subspace. In Section 3 we discuss the closeness of the range of bounded linear operator F_AB:X→AX-XB, and prove that R(δ_A)∩{A}′∩{Aⁿ}′=0, where δ_A:X→AX-XA.     

The monodromy of a series of hypersurface singularities

Let {f=0} be a hypersurface inC ⁿ⁺¹ with a 1-dimensional singular set Σ. We consider the series of hypersurfaces {f+ɛx ^N=0} wherex is a generic linear form. We derive a formula, which relates the characteristic polynomials of the monodromies off andf+ɛx ^N. Other ingredients in this formula are the horizontal and the vertical monodromies of the transversal (isolated) singularities on each branch of the singular set. We use polar curves and the carrousel method in the proof. The formula is a generalization of the Iomdin formula for the Milnor numbers: μ(f+ɛx ^N )=μ _n (f)−μ _n −1(f)+Ne ₀(Σ)     

线性常微分方程组解的特征数的估

复数域上线性系统$\dot x=A(t)x$     

On Two-stage Estimate Based on Independent Estimate of Covariance Matrix

When an independent estimate of covariance matrix is available, we often prefer two-stage estimate （TSE）. Expressions of exact covarianee matrix of the TSE obtained by using all and some covariables in eovariance adjustment approach are given, and a necessary and sufficient condition for the TSE to be superior to the least square estimate and related large sample test is also established. Furthermore the TSE, by using some covariables, is expressed as weighted least square estimate. Basing on this fact, a necessary and sufficient condition for the TSE by using some covariables to be superior to the TSE by using all eovariables is obtained. These results give us some insight into the selection of covariables in the TSE and its application.     

Estimate for stochastic integrals

Lithuanian Mathematical Journal -     

基于泛岭估计对岭估计过度压缩的改进方法

岭估计是解决多元线性回归多重共线性问题的有效方法,是有偏的压缩估计。与普通最小二乘估计相比,岭估计可以降低参数估计的均方误差,但是却增大残差平方和,拟合效果变差。本文提出一种基于泛岭估计对岭估计过度压缩的改进方法,可以改进岭估计的拟合效果,减小岭估计残差平方和的增加幅度。     

Logistic模型中参数的估计

本文得到Ｌｏｇｉｓｔｉｃ模型中参数ａｉ、ｂｉ 的一种便于使用的估计公式,在样本容量较小及精度要求不高时用作试题参数的估计是合适的,也可作为联合极大似然估计中的初始值     

随机估计和VDR检验

众所周知统计推断有三种理论:普遍承认的Neyman理论(频率学派),Bayes推断和信仰推断(Fiducial)。Bayes推断基于后验分布,由先验分布和样本分布求得。信仰推断是基于信仰分布(Confidence Distribution,简称CD),直接利用样本求得。两者推断方式一致,都是用分布函数作推断,称为分布推断。从分析传统的参数估计、假设检验特性来看,经典统计推断也可以视为分布推断。通常将置信上限看做置信度的函数。其反函数,即置信度是置信上界的函数,恰是分布函数,该分布恰是近年来引起许多学者兴趣的CD。在本文中,基于随机化估计(其分布是一CD)的概率密度函数,提出VDR检验。常见正态分布期望或方差的检验,多元正态分布期望的Hoteling检验等是其特例。VDR(vertical density representation)检验适合于多元分布参数检验,实现了非正态的多元线性变换分布族的参数检验。VDR构造的参数的置信域有最小Lebesgue测度。     

制导参量的估计方法

为了对机动目标实施有效的垂直命中攻击,必须对制导参量进行估计．作者根据实际战斗条件下目标机动特性以及导出参量的噪声性质,对有色噪声进行了白化处理,并给出了机动目标速度、加速度的自适应滤波算法．该方法可有效提高制导参量的估计精度．     

On Li–Yau Heat Kernel Estimate

We present some improvements of the Li–Yau heat kernel estimate on a Riemannian manifold with Ricci curvature bounded below. As a consequence we prove a gradient estimate to the heat kernel with an optimal leading term.     

Bayes估计的哲学思考

从哲学角度讨论了参数的Bayes估计,对经典学派与Bayes学派的分歧进行了一些哲学思考,这实际上是唯物主义与唯心主义的分歧.作者认为唯物论与唯心论只是认识世界的两个观点,它统一与客观世界.     

The Hilbert Kernel Regression Estimate

Let (X, Y) be an ^d× -valued regression pair, whereXhas a density andYis bounded. Ifni.i.d. samples are drawn from this distribution, the Nadaraya–Watson kernel regression estimate in ^dwith Hilbert kernelK(x)=1/x^dis shown to converge weakly for all such regression pairs. We also show that strong convergence cannot be obtained. This is particularly interesting as this regression estimate does not have a smoothing parameter.