一般δ-冲击模型中无失效数据的Bayes统计推断

在这篇文章中,我们针对一般冲击模型,研究Bayes方法处理无失效数据的问题.所谓一般δ-冲击模型是指系统受到强度为λ的Poisson冲击,当两个连续冲击之间时间间隔的长度不属于某个固定的区间[δ1,δ2]时,系统将失效.我们分别选择均匀分布和Beta分布作为先验分布,用Bayes方法和多层Bayes方法得到了参数δ1和δ2的估计.     

���ڿ�ζ�Ƚϵļ���

本文用Bayes方法来检验两种药品之间的口味是否相同, 得到了与人们的直观感觉相吻合的结论\bd 设$(x_1,x_2,x_3)$服从参数为$(p_1,p_2,p_3)$的多项分布, 则检验口味的问题转为检验$H_i:(p_1,p_2,p_3)$ $\in\Theta_i\;\;i=1,2,3$三个假设的问题. 在使用Bayes方法时, 其关键是选择合理的先验分布, 这里选择相应的先验分布为无信息先验, 据此求出每个假设的后验概率, 从而得到结论\bd     

指数分布参数多层Bayes和E Bayes估计的性质

本文讨论无失效数据下指数分布参数多层Bayes估计和E Bayes估计的性质,在超参数分别取两种不同的先验分布下,证明参数的多层Bayes估计和E Bayes估计渐近相等,且多层Bayes估计值小于E Bayes估计值.     

无失效数据的Bayes和多层Bayes估计

对无失效数据的研究 ,是近些年来遇到的一个新问题 ,在实际问题中迫切需要解决 ,这项工作具有理论和实际应用价值 .本文对无失效数据 (ti,ni) ,在时刻ti 的失效概率pi=p{T 相似文献   

二项分布参数多层Bayes和E Bayes估计的性质

讨论无失效数据下二项分布参数E Bayes估计和多层Bayes估计的性质,证明二项参数的多层Bayes估计和E Bayes估计渐近相等,且E Bayes估计值小于多层Bayes估计值.     

无失效数据失效概率的Bayes估计

本文对无失效数据(ti,ni)在ti 时刻的失效概率pi= P｛T< ti｝的先验密度的核为(1- pi)k时,给出了pi 的Bayes估计和多层Bayes估计,由此可以得到无失效数据可靠度的估计.最后,结合实际问题进行了计算.     

逐步增加首失效截尾样本下参数估计的优良性

在对称平方损失函数下, 利用逐步增加首失效截尾样本, 研究两参数Pareto分布族参数的一致最小方差无偏估计(UMVUE), Bayes估计和参数型经验Bayes(PEB)估计. 按照均方误差(MSE)准则, 比较UMVUE与PEB估计的优良性. 根据风险函数导出Bayes估计与PEB估计的渐近性, 并获得它们的收敛速度. 在相同的置信水平下, 研究参数分别在经典统计和Bayes统计中的区间估计, 并利用数值模拟说明Bayes区间估计的精度高于经典统计区间估计.     

无失效数据情形参数的综合估计

本对指数分布的无失效数据,在引进失效信息后,在先验分布为Gamma分布时,给出了失效率的多层Bayes估计和综合Bayes估计,并给出了无失效数据情形可靠度的综合估计,还结合实际问题进行了计算。     

不完全椭球约束下多元线性模型线性估计的可容许性

本文研究了多元线性模型当未知参数受不完全椭球约束$\mbox{tr}(\Theta-\Theta_1)'N(\Theta-\Theta_1)\leq\sigma^2$时线性估计的可容许性问题.具体而言,我们研究了约束$\mbox{tr}(\Theta-\Theta_1)'N(\Theta-\Theta_1)\leq\sigma^2$中$N$和非中心点$\Theta_1$对线性估计的可容许性的影响.主要结果表明在两个不同的不完全椭球约束条件$\mbox{tr}(\Theta-\Theta_1)'N(\Theta-\Theta_1)\leq\sigma^2$与$\mbox{tr}(\Theta-\Theta_2)'N(\Theta-\Theta_2)\leq\sigma^2$ 下,当$\Theta_1$和$\Theta_2$满足一定的关系时,可容许的齐次线性估计类是相同的.     

产品失效率的验前分布稳健性分析

研究Bayes统计分析中利用验前信息的稳健性.首先,用一般方法研究了指数寿命型分布中失效率的验前分布的稳健性.然后利用Gamma分布函数的典型性质,并以平方损失下的后验期望损失为判别准则,讨论了失效率的最优Bayes稳健区间.给出了失效率的最优Bayes稳健点估计.     

由圈长分布确定的二部图~$K_{n,n r}-A\ (|A|\leq 3)$

The cycle length distribution of a graph G of order n is a sequence （c1 （G）,…, cn （G））, where ci （G） is the number of cycles of length i in G. In general, the graphs with cycle length distribution （c1（G） ,…,cn（G）） are not unique. A graph G is determined by its cycle length distribution if the graph with cycle length distribution （c1 （G）,…, cn （G）） is unique. Let Kn,n＋r be a complete bipartite graph and A lohtaib in E（Kn,n＋r）. In this paper, we obtain： Let s 〉 1 be an integer. （1） If r = 2s, n 〉 s（s - 1） ＋ 2｜A｜, then Kn,n＋r - A （A lohtain in E（Kn,n＋r）,｜A｜ ≤ 3） is determined by its cycle length distribution; （2） If r = 2s ＋ 1,n 〉 s^2 ＋ 2｜A｜, Kn,n＋r - A （A lohtain in E（Kn,n＋r）, ｜A｜ ≤3） is determined by its cycle length distribution.     

On Hermite-Hadamard-type characterizations of higher-order differential inequalities

Let $I$ be an open interval of $\mathbb{R}$ and $f: I\to \mathbb{R}$. It is well-known that $f$ is convex in $I$ if and only if, for all $x,y\in I$ with $x相似文献   

ON ONE CONJECTURE OF R. S. SINGH

In 1979 R. S. Singh(Ann. Statist, 1979, p. 890) made a conjecture concerning the convergence rate of EB estimates of the parameter θ in an one-dimensional continuous exponential distribution family, under the square loss function, the prior distribution family being confined to a bounded interval. The conjecture asserts that the rate cannot reach o(1/n) or even O(1/n). In this article, the weaker part of this conjecture(i. e. the o(1/n) part) is shown to be correct.     

DIRECT EXPANSIONS FOR THE DISTRIBUTION FUNCTIONS AND DENSITY FUNCTIONS OF χ2-TYPE AND t-TYPE DISTRIBUTED RANDOM VARIABLES

Suppose that Z1,Z2…,Zn are independent normal random variables with common mean μ and variance σ^2. Then S^2=∑n n=1 (zi-z)^2/σ^2 and T =（n-1的平方根）-Z/（S^2/n的平方根） have x2n-1 distribution and tn-1 distribution respectively. If the normal assumption fails, there will be the remainders of the distribution functions and density functions. This paper gives the direct expansions of distribution functions and density functions of S^2 and T up to o(n^-1). They are more intuitive and convenient than usual Edgeworth expansions.     

广义线性模型中检验回归参数统计量的渐近分布

在广义线性模型中,若(对某个α>0),且其它一些正则条件满足,可以证明Wald检验统计量的渐近分布是X2分布,其中,是ZiZi'的最小特征根,Zi是有界的p×q回归系数,yi是q×1响应变量．     

负相协随机变量非随机和的差的精确大偏差

In this paper,we study precise large deviation for the non-random difference sum from j=1 to n_1(t) X_(1j)-sum from j=1 to n_2(t) X_(2j),where sum from j=1 to n_1(t) X_(1j) is the non-random sum of {X_(1j),j≥1} which is a sequence of negatively associated random variables with common distribution F_1(x),and sum from j=1 to n_2(t) X_(2j) is the non-random sum of {X_(2j),j≥1} which is a sequence of independent and identically distributed random variables,n_1(t) and n_2(t) are two positive integer functions.Under some other mild conditions,we establish the following uniformly asymptotic relation lim t→∞ sup x≥r(n_1(t))~(p+1)|(P(∑~(n_1(t)_(j=1)X_(1j)-∑~(n_2(t)_(j=1)X_(2j)-(μ_1n_1(t)-μ_2n_2(t)x))/(n_1(t)F_1(x))-1|=0.     

Absolutely continuous spectrum for random Schrödinger operators on the Bethe strip

The Bethe strip of width m is the cartesian product $\mathbb {B}\times \lbrace 1,\ldots ,m\rbrace$, where $\mathbb {B}$ is the Bethe lattice (Cayley tree). We prove that Anderson models on the Bethe strip have “extended states” for small disorder. More precisely, we consider Anderson‐like Hamiltonians $H_\lambda =\frac{1}{2} \Delta \otimes 1 + 1 \otimes A\,+\,\lambda \mathcal {V}$ on a Bethe strip with connectivity K ≥ 2, where A is an m × m symmetric matrix, $\mathcal {V}$ is a random matrix potential, and λ is the disorder parameter. Given any closed interval $I\subset \big (\!-\!\sqrt{K}+a_{{\rm max}},\sqrt{K}+a_{\rm {min}}\big )$, where a_min and a_max are the smallest and largest eigenvalues of the matrix A, we prove that for λ small the random Schrödinger operator H_λ has purely absolutely continuous spectrum in I with probability one and its integrated density of states is continuously differentiable on the interval I.     

Fractional Hamiltonian systems with positive semi-definite matrix

We study the existence of solutions for the following fractional Hamiltonian systems $$ \left\{ \begin{array}{ll} - _tD^{\alpha}_{\infty}(_{-\infty}D^{\alpha}_{t}u(t))-\lambda L(t)u(t)+\nabla W(t,u(t))=0,\\[0.1cm] u\in H^{\alpha}(\mathbb{R},\mathbb{R}^n), \end{array} \right. ~~~~~~~~~~~~~~~~~(FHS)_\lambda $$ where $\alpha\in (1/2,1)$, $t\in \mathbb{R}$, $u\in \mathbb{R}^n$, $\lambda>0$ is a parameter, $L\in C(\mathbb{R},\mathbb{R}^{n^2})$ is a symmetric matrix, $W\in C^1(\mathbb{R} \times \mathbb{R}^n,\mathbb{R})$. Assuming that $L(t)$ is a positive semi-definite symmetric matrix, that is, $L(t)\equiv 0$ is allowed to occur in some finite interval $T$ of $\mathbb{R}$, $W(t,u)$ satisfies some superquadratic conditions weaker than Ambrosetti-Rabinowitz condition, we show that (FHS)$_\lambda$ has a solution which vanishes on $\mathbb{R}\setminus T$ as $\lambda \to \infty$, and converges to some $\tilde{u}\in H^{\alpha}(\R, \R^n)$. Here, $\tilde{u}\in E_{0}^{\alpha}$ is a solution of the Dirichlet BVP for fractional systems on the finite interval $T$. Our results are new and improve recent results in the literature even in the case $\alpha =1$.