非参数估计

First, we obtain the empirical Bayesian estimator of the scale parameter for the double exponential distribution based on LINEX loss function and the convergence rate of the estimate.

首先利用非对称的LINEX损失函数对双指数分布族刻度参数进行了经验贝叶斯估计，并讨论了该估计的性质，给出了收敛速度。

We make the following assumption for When 2 is positive definite matrix, different estimators about matrix of regression coefficients and inefficiency of Least squares estimate have been discussed in many documents. Considered 2 is nonnegative definite matrix, this thesis derives Best linear unbiased Estimate of parameter matrix B and estimable parameter function KBL under the meaning of matrix nonnegative definite and the property of maximum probability of BLUE is investigated.

当∑＞0时，众多文献讨论了回归系数阵的各种估计及LSE的有效性，本文考虑了当∑≥0的情形，给出了回归系数阵B及其可估参数函数KBL的在矩阵非负定意义下的最优估计，研究了它的一个最大概率性质，并且讨论了最小二乘估计成为最佳线性无偏估计的充分必要条件，在此基础上给出了均值矩阵的最小二乘估计与BLUE的偏差估计，定义了LSE相对于BLUE的一个相对效率，并给出了它的界。

In the case of NA samples, the parameter for scale exponential family is discussed using empi- rical Bayes methods. Moreover, we investigate the statistical analysis of life data in reliability tests. Firstly, based on the LINEX loss function, we obtain the empirical Bayesian estimator for parameter of the one-side truncated-type distribution families and the scale exponential family under NA samples, also discuss the estimators?

在非对称的LINEX损失函数下，我们首先给出了一类单边截断型分布族、刻度指数族参数的经验Bayes估计NA样本情形），并讨论了该估计的收敛性质，在某些条件下，证明了本文给出的经验Bayes估计是渐近最优的，且给出了收敛速度。

Combining NA samples the EB estimation with convergence rates is also obtained; Thirdly, for one-parameter exponential distribution family, we give the parameters' EB estimator that is admissible and asymptotically optimal with convergence rates. Finally, Bayesian and hierarchical Bayesian approaches are applied to analyze the reliability performances for series system with cold standby units and the numerical simulation results show that multiple Bayes estimator is superior to Bayes estimator.

其次，针对一类双边截断分布族，在非对称linex损失下构造了经验Bayes决策函数，建立了它的收敛速度，给出了渐近最优的证明；并结合NA样本研究了参数的经验Bayes估计及其收敛速度；接着给出了指数族参数的经验Bayes估计，证明了它是渐近最优的，也是可容许的，同时也获得了该EB估计的收敛速度；最后，运用Bayes与多层Bayes方法研究了具有冷贮备部件串联系统的可靠性指标的估计，并给出了数值模拟结果，结果表明多层Bayes估计优于Bayes估计。

Generalized linear models, which can model a large variety of data, have a wide area of application. The class of GLMs includes, as special cases, linear regression, analysis-of-variance models, log-linear models for the analysis of contingency tables, logistic models for binary data in the form of proportions and many others. Usually, the parameters in the generalized linear models are estimated by the method of maximum likelihood . But, in the literature, the nonrobustness of the maximum likelihood estimator forβhas been studied extensively. The quasi-likelihood estimator of the parameter of the generalized linear model shares the same non-robustness properties.

广义线性模型，可用于对多种类型的数据进行建模，是应用非常广泛的模型，线性回归模型、方差分析模型、用于列联表分析的对数线性模型和逻辑斯谛模型等都是广义线性模型的特例，通常，我们用极大似然的方法估计广义线性模型中的参数，但是，在文献中，对参数β的极大似然估计的非稳健性已经有了广泛的研究，广义线性模型的拟似然估计也显示了非稳健性。

The experiment results prove that the way with regularization technology is better than the way without regularization technology. In order to overcome the limitation of the regularization method, chapter 6 builds an irregular MRF model suitable for the motion estimation of feature point based on MAP-MRF frame. The energy function, which can reflect the joint probability distribution of motion parameters and the constraints relation of local 3D motion parameters, is established in irregular MRF.

为了解决正则化方法只能分析非刚体局部三维运动的缺点，第6章在基于MAP-MRF的分析框架中建立了与特征点运动估计相对应的非规则MRF模型，构造了非规则MRF模型中反映非刚体局部三维运动参数之间约束关系和运动参数联合概率分布的能量函数，通过能量函数最小化估计运动参数。

For noncentrality parameter matrix in noncentral wishart distribution, based on usual unbiased estimator, a series of pre-test estimators are obtained.

1对于非中心Wishart分布的非中心参数矩阵，本文基于通常的无偏估计给出了一系列的预检验估计。

Two kinds of DPA estimators are given by way of Nonstandard Analysis andprobability parameter sampling,respectively.One kind of DPA algorithm,based on a sample pathof GI/GI/l system,can estimate simultaneously EST and expected busy period length gradient.

分别从非标准分析和概率参数抽样方式的角度建立了两类DPA估计量，特别地，前一类DPA估计量可以同时估计顾客的期望系统时间和忙期期望长度的梯度。

By time-varying filtering, we realize the signal selectivity and interference suppression, especially for the signals with time-varying power spectrum. It is shown herein that conventional eigenstructure based methods are lower dimensional special case of the time-frequency subspace fitting method. The closed-form array response estimation algorithms for blind source separation in wireless communication is also developed in the dissertation by exploiting the data structure of second-order statistics of the array output in the presence of multipath. The proposed method achieves array response estimation with little constraint on signal property and propagation environment such as scatters or angular spread. Moreover, the array utilized here can be of arbitrary geometry and even uncalibrated. Since the closed-form solution is obtained, no misconvergency may occur and large amount of computation can be saved. The equivalent feed model and corresponding optimization designation are proposed to realize the downlink multi-user selective transmission. By coherent and non-coherent superposition of the electromagnetical field, multiple beams can be formed using a single array, so that each co-channel user signal is transmitted in each beam and will not interfered with each other. The proposed system scheme implements the digital beamforming and the channel allocation in the base band, therefore is suitable for the utilization of the radio software technology. The feasibility and effectiveness of the proposed model are proved by the RF experiment.

在窄带条件下，GSC－SSF方法还可以提高阵列的虚拟孔径，从而提高估计的性能；可以证明，Cyclic DOA 和SC-SSF方法可看作该方法的特例或近似； 3、提出了二维GSC－SSF方法和宽带谱相关DOA矩阵方法，利用信号的循环平稳特性解决了宽带条件下二维 DOA 估计问题，并具有阵元数少、计算量较小的特点； 4、提出了时频子空间拟合和宽带时频DOA估计算法，将联合时频分析的理论和方法引入到阵列信号的空间谱估计研究中，利用时变滤波实现信号选择性测向和干扰抑制，特别适合于谱时变信号；可以证明传统特征结构法可看作是时频子空间拟合方法的特例； 5、提出了多用户环境下阵列响应的估计方法——CARE方法，该方法对信号结构和传播参数（主要是散射源分布与角度散布）的约束较少，适用于各种几何结构的阵列甚至未校准的阵列，能够直接得到阵列响应的闭式解，因此计算量小、稳健性好，不存在收敛问题； 6、提出了针对多用户下行选择性发送的阵列等效激励模型、优化设计方法及其系统实现方案，通过空间场的相干叠加与非相干叠加使得单一阵列能够形成多个波束，并保证各波束内的用户信号各不相同且相互干扰最小；其系统方案具有在基带实现数字波束成形、信道分配并便于软件无线电实现的特点，其有效性已得到外场实测数据的证实。

Objective:This paper proposes an iterative noncentral procedure for the sample size determination in log rank test and compares it with the LachinFoulkes procedure. Methods: Samplings are performed with a prescribed size.

本研究首先按临床数据实际分析过程说明设计参数赋值依据和生存研究所需样本量测定原则，然后提出迭代非中心法，即借助重复抽样估计其非中心参数，测定所需样本量，并同基于指数分布的方法进行比较。

With Death guitarist Schuldiner adopting vocal duties, the band made a major impact on the scene.

随着死亡的吉他手Schuldiner接受主唱的职务，乐队在现实中树立了重要的影响。

But he could still end up breakfasting on Swiss-government issue muesli because all six are accused of nicking around 45 million pounds they should have paid to FIFA.

不过他最后仍有可能沦为瑞士政府&议事餐桌&上的一道早餐，因为这所有六个人都被指控把本应支付给国际足联的大约4500万英镑骗了个精光。