非参数回归函数的基于截尾数据的估计

本文考虑截尾数据情况下非参数回归函数m(x)=E(Y|x)的估计。具体地讲,我们面对的是这样的数学模型:T是与(X,Y)独立的随机变量,我们观测到的不是Y本身,而是Z=min(Y,T)及δ=[Y≤T]。今有训练样本{(X_i,Z_i,δ_i)}_(i-1)及当前样本(X,z,δ),记ξ_i(·)=[z_i≥·], N~ (·)=sum from i=1 to n ξ_i(·), V_n(·)=multiply from i=1 to n{1 N~ (z_i)/2 N~ (z_i)}~[δ_i=_i<0], U_n(·)=sum from i=1 to n Wnt(x)ξ_i(·), 令 m_n(x)=integral from 0 to u_n U_n(y)|V_n(y)dy, 其中u_n=F_2~(-1)(n~(-a)),0<α<1/2为一实常数,F_2(·)=P(Y≥·)为Y的(右侧)分布函数。在权函数{W_(ni)(x)}_(i=1)~n及(X,Y,T)的分布函数满足一组条件下,我们证明了m_n(x)为m(x)的强相合估计,即:m_n(x)→m(x),a.s.(n→ ∞).     

Richard模型的平均期望费用问题

设 W_t,t≥0为(Ω,■,P)上标准 Wiener 过程,■为由之所生成的上升 σ-域族,以τ_i,i≥1表任一个(?)单调上升停时列,对每个τ_i 确定一个F_(τi)可测随机变量ξ_i,我们称任一这样的对列 v={(τ_i,ξ_i),i≥1}为一脉冲过程,以 V 表脉冲过程,(以下称脉冲控制)的全体,设 h 和 B 为 R 上满足某些条件的非负实函数,再设σ,μ为任何实常数且|σ|>0.本文求得一个常数λ>0使对任一实数 x 皆有一个十分明确具体的控制 v~*={(τ_i~*,ξ_i~*),i≥1)∈V 满足lim~T→∞(1/T)E[integral from 0 to T h(x+μt+σW_t+sum τ_i~*相似文献   

关于B值随机元赋范部分和最大值的矩的若干讨论

本文讨论B值随机元部分和序列的最大值的矩的问题,对1≤p≤2及r>p证明了下列叙述的等价性; (ⅰ)存在常数0相似文献   

On Diophantine Approximation and Approximate Analysis (Ⅰ)

<正> We use γ=(γ_(11),…γ_(t1),…γ_(15)…,γ_(ts))to denote a point in st-dimensionalEuclidean space R_(st).We use the notations γ_i=(γ_(li)…γ_(ti)(1≤i≤s)and γ_i=(γ_i1,…,γ_(is))(1≤j≤t).q=(q_1,…,q_t),k=(k_1,…,k_s)and m=(m_1,…,m_s)will be     

关于Littlewood的一个问题

本文证明了: (1)如果{a_n}_n~N=1是非负不减序列,p>0,q>0,0≤r≤1,且p(q+r)≥q+p,则sum from n=1 to N(a_n~pA_n~q)(sum from m=n to N(a_n~(1+p/q)~r≤1·sum from n=1 to N(a_n~pA_n~q)~(1+p/q),其中A_n=sum from m=n to n (a_m).上述不等式在0≤r≤1时完全解决了H.Alzer~([4])在1996年提出的一个问题,且1是最佳常数; (2)如果{a_n}_n~N=1是非负序列,p,p≥1,r>0,r(p-1)≤2(q-1),令α=((p-1)(q+r)+p~2+1)/(p+1) β=(2p+2r+p-1)/(q+1),σ=(q+r-1)/(p+q+r)则sum from n=1 to N (a_n~p)sum from i=1 to n (a_i~qA_i~r)≤2~σsum from n=1 to N(a_n~αA_n~β)(0.2)(0.2)式改进了G.Be(?)et~([2,3])在1987年对Littlewood一个问题的结果,常数因子的3/2降为2~(3/2)=1.2598…     

关于误差方差估计渐近正态收敛速度的上,下界

考虑线性回归模型 Y_■=x_4~′β+e_■ i=1,2,…设误差序列■,i≥1满足条件:e_■ i≥1 i.i.d.,Ee_1=0,Ee_1~2=σ~2>0,∞>Var e_1~2=τ~2>0。记■_n~2=1/(n-r){sum from j=1 to n e■-sum from k=1 to r (sum from j=1 to n a_(akj)■_j)~2} δ(n)=τ~(-2)E(■_1~2-σ~2)~2I_((|■-σ~2|≥■τ)+τ~(-3)n~(1/2)|E(■_1~2-σ~2)~3I_((|■_1~2-σ~2|<(nτ)~(1/2))+τ~(-4)n~(-1)E■_1~2-σ~2)~4I_((|■-σ~2|0使得■|P(■_n~2-σ~2)/(Var■_n~2)~(1/2))≤x)-Φ(x)|≤C(δ(n)+n~(-1/2)) ■|P(■_n~2-σ~2)/(Var■_n~2)~(1/2))≤x)-Φ(x)|+n~(-1/2)≥C_1δ(n)。     

二次指派问题的一个新的限界方法

二次指派问题(QAP)的数学模型是:min{z(x)=sum from i=1 to n sum from =1 to n a_(ip)x_(ip)+sum from i=1 to n sum from p=1 to n sum from j=1 to n sum from q=1 to n c_(ipjq)x_(ip)x_(jq)|x∈},(1)这里∈(n~2维布尔集)是满足如下约束的集合:sum from i=1 to n x_(ip)=1,1≤p≤n,(2)sum from p=1 to n x_(ip)=1,1≤i≤n,(3)x_(ip)=0,1,1≤i,p≤n.(4)因为 x_(ip)~2=x_(ip)并且有约束(2)和(3),我们可以约定 c_(ipjq)=0,当 i=j 或 p=q.如果所有二次项的系数都可以写成     

回归函数基于分割的改良递推估计的渐近正态性

提出了回归函数基于分割的改良递推估计,mn(x)=sum (IAn(x) (Xi)YiI(|Yi|) from i=1 to n≤bi)/sum (IAn(x)(Xi)) from i=1 to n,并在较为简洁的条件下证明了该估计量的渐近正态性.     

亚纯单叶函数的Hadamard乘积

设g(ζ)=ζ+sum from n=0 to ∞b_nζ~(-n)为α级亚纯星形函数(0≤α<1,|ζ|>1),函数ψ(z)=z+sum from n=2 to ∞α_nZ~n为单位园内的凸单叶函数。本文得到,若α∈(1/2,1),则g(ζ)※ζ~2ψ(1/ζ)(|ζ|>1)为α级亚纯星形函数,作为这个结果的一个推论,文[4]中的猜测在α∈(1/2,1)内成立。     

关于线性过程谱函数估计的弱不变原理的一点注记

设随机序列{X_n; n=0,±1…}可表示成为X_n=sum from j=-∞ to +∞(α_(j-n)ζ_j其中{α_j}是满足sum from j=-∞ to +∞(α_j~2)<∞的实数列,{ζ_j}是白噪声序列。通常用(?)_N(λ)=integral from 0 to λ(1/2πN)∣sum from k=1 to N(x_(?)e~(iμk)∣~2 dμ来估计{x_n}的未知的谱函数F(λ)。在一定的条件下,当{ζ_j}是独立同分布随机序列时,和[3]证明了:过程√(?)[(?)_N(λ)-F(λ)]的分布弱收敛到某个正态过程ζ(λ)在C[0,π]上产生的测度。本文在他们工作的基础上,运用鞅的极限定理和鞅不等式,改进了[3]中的两个关键引理,从而证明了当{ζ_j}是有控制分布的实四阶鞅差序列时,仍有相同的结果。     

Small divisors of Bernoulli sums

Let ?= {?_i,i ≥1} be a sequence of independent Bernoulli random variables (P{?_i = 0} = P{?_i = 1 } = 1/2) with basic probability space (Ω, A, P). Consider the sequence of partial sums B_n=?₁+...+?_n, n=1,2..... We obtain an asymptotic estimate for the probability P{P^-(B_n) > >} for >≤n^{e/log log n}, c a positive constant.     

Stability of Maxima of Random Variables with Multidimensional Indices

The paper discusses the stability of suitably-defined maxima of a set of i.i.d. random variables with multidimensional indices.It is shown that theorems of Gnedenko (1943) and Tomkins (1986) concerning relative stability and complete relative stability of maxima remain valid in the new setting.Moreover, a criterion for almost sure relative stability for maxima with multidimensional indices is presented, extending a result of Barndorff-Nielsen (1963).AMS 2000 Subject Classification. Primary—60F15, 60G60, 62G30, Secondary—10A25, 60G99     

Small deviation probabilities of weighted sums under minimal moment assumptions

We examine small deviation probabilities of weighted sums of i.i.d.r.v. with a power decay at zero under moment assumptions close to necessary.     

Complete moment convergence for i.i.d. random variables

In the paper, the complete moment convergence is obtained for i.i.d. random variables such that all moments exist, but the moment generating function does not exist. The main results extend the related known works due to Gut and Stadtmüller.     

Saddlepoint approximation for moments of random variables

In this paper, we introduce a saddlepoint approximation method for higher-order moments like E(S − a)₊ ^m , a>0, where the random variable S in these expectations could be a single random variable as well as the average or sum of some i.i.d random variables, and a > 0 is a constant. Numerical results are given to show the accuracy of this approximation method.     

Attainable rates of convergence of maxima

Any exponential rate of convergence can be obtained for maxima of i.i.d. random variables, while faster than exponential convergence implies that the variables have an extreme value distribution.     

Precise rates in the law of logarithm for the moment convergence of i.i.d. random variables

Let be a sequence of i.i.d. random variables with EX=0 and EX²=σ²<∞. Set , Mn=maxk?n|Sk|, n?1. Let r>1, then we obtain

     

Precise asymptotics for a series of T. L. Lai

Let be i.i.d. random variables with , and set . We prove that, for

under the assumption that and Necessary and sufficient conditions for the convergence of the sum above were established by Lai (1974).