Exponential inequalities for associated random variables and strong laws of large numbers

Some exponential inequalities for partial sums of associated random variables are established. These inequalities improve the corresponding results obtained by Ioannides and Roussas (1999), and Oliveira (2005). As application, some strong laws of large numbers are given. For the case of geometrically decreasing covariances, we obtain the rate of convergence n-1/2(log log n)1/2(logn) which is close to the optimal achievable convergence rate for independent random variables under an iterated logarithm, while Ioannides and Roussas (1999), and Oliveira (2005) only got n-1/3(logn)2/3 and n-1/3(logn)5/3, separately.     

Empirical Bayes Test for the Parameter of Rayleigh Distribution with Error of Measurement

For the data with error of measurement in historical samples, the empirical Bayes test rule for the parameter of Rayleigh distribution is constructed, and the asymptotically optimal property is obtained. It is shown that the convergence rate of the proposed EB test rule can be arbitrarily close to O（n-1/2） under suitable conditions.     

Local and global convergence of a general inertial proximal splitting scheme for minimizing composite functions

This paper is concerned with convex composite minimization problems in a Hilbert space. In these problems, the objective is the sum of two closed, proper, and convex functions where one is smooth and the other admits a computationally inexpensive proximal operator. We analyze a family of generalized inertial proximal splitting algorithms (GIPSA) for solving such problems. We establish weak convergence of the generated sequence when the minimum is attained. Our analysis unifies and extends several previous results. We then focus on \(\ell _1\)-regularized optimization, which is the ubiquitous special case where the nonsmooth term is the \(\ell _1\)-norm. For certain parameter choices, GIPSA is amenable to a local analysis for this problem. For these choices we show that GIPSA achieves finite “active manifold identification”, i.e. convergence in a finite number of iterations to the optimal support and sign, after which GIPSA reduces to minimizing a local smooth function. We prove local linear convergence under either restricted strong convexity or a strict complementarity condition. We determine the rate in terms of the inertia, stepsize, and local curvature. Our local analysis is applicable to certain recent variants of the Fast Iterative Shrinkage–Thresholding Algorithm (FISTA), for which we establish active manifold identification and local linear convergence. Based on our analysis we propose a momentum restart scheme in these FISTA variants to obtain the optimal local linear convergence rate while maintaining desirable global properties.     

非线性模型误差方差估计的Bootstrap逼近和分布的收敛速度

本文对非线性模型误差方差的估计基于Jackknife虚拟值的Bootstrap方法建立了Bootstrap逼近,证明了逼近的相合性定理,得到了逼近的速度是o(n~(-1/2))。进一步,本文证明了误差方差估计的分布以理想的最佳速度o(n~(-1/2))收敛于正态分布的结论。     

The Empirical Bayes Estimation of the Location Parameter in Double-Exponential Family under the LINEX Loss Function

In this paper, we derive the Bayes estimator of the location parameter in double-exponential family under the LINEX loss function, and then construct the corresponding empirical Bayes estimator. It is shown that the empirical Bayes estimator is asymptotically optimal with convergence rate being , , where 1/2相似文献   

多维广义线性模型拟极大似然估计的弱相合性

本文考虑多维广义线性模型的拟似然方程$\tsm^n_{i=1}X_i(y_i-\mu(X_i'\xb))=0$, 在一定条件下证明了此方程的解$\wh\xb_n$渐近存在, 并得到了其收敛速度, 即$\wh\xb_n-\xb_0=O_p({\underline{\xl}}_n^{-1/2})$, 其中$\xb_0$为参数$\xb$的真值, $\underline{\xl}_n$是方阵$S_n=\tsm^n_{i=1}X_iX_i'$的最小特征值。     

完全随机缺失机制下伽玛分布中形状参数的经验贝叶斯推断

在完全随机缺失机制下构造了伽玛分布参数的经验贝叶斯检验函数,并获得了它的渐进最优性.在适当的条件下证明了所提出的经验贝叶斯检验函数收敛速度可任意接近O(n(~(-1/2)).     

Optimization of the Hermitian and Skew-Hermitian Splitting Iteration for Saddle-Point Problems

We study the asymptotic rate of convergence of the alternating Hermitian/skew-Hermitian iteration for solving saddle-point problems arising in the discretization of elliptic partial differential equations. By a careful analysis of the iterative scheme at the continuous level we determine optimal convergence parameters for the model problem of the Poisson equation written in div-grad form. We show that the optimized convergence rate for small mesh parameter h is asymptotically 1–O(h ^1/2). Furthermore we show that when the splitting is used as a preconditioner for a Krylov method, a different optimization leading to two clusters in the spectrum gives an optimal, h-independent, convergence rate. The theoretical analysis is supported by numerical experiments.This revised version was published online in October 2005 with corrections to the Cover Date.     

Aggregated estimators and empirical complexity for least square regression

Numerous empirical results have shown that combining regression procedures can be a very efficient method. This work provides PAC bounds for the L² generalization error of such methods. The interest of these bounds are twofold.First, it gives for any aggregating procedure a bound for the expected risk depending on the empirical risk and the empirical complexity measured by the Kullback–Leibler divergence between the aggregating distribution and a prior distribution π and by the empirical mean of the variance of the regression functions under the probability .Secondly, by structural risk minimization, we derive an aggregating procedure which takes advantage of the unknown properties of the best mixture : when the best convex combination of d regression functions belongs to the d initial functions (i.e. when combining does not make the bias decrease), the convergence rate is of order (logd)/N. In the worst case, our combining procedure achieves a convergence rate of order which is known to be optimal in a uniform sense when (see [A. Nemirovski, in: Probability Summer School, Saint Flour, 1998; Y. Yang, Aggregating regression procedures for a better performance, 2001]).As in AdaBoost, our aggregating distribution tends to favor functions which disagree with the mixture on mispredicted points. Our algorithm is tested on artificial classification data (which have been also used for testing other boosting methods, such as AdaBoost).     

求解定常不可压Stokes方程的两层罚函数方法

借助于两套有限元网格空间提出了一种求解定常不可压Stokes方程的两层罚函数方法.该方法只需要求解粗网格空间上的Stokes方程和细网格空间上的两个易于求解的罚参数方程（离散后的线性方程组具有相同的对称正定系数矩阵）.收敛性分析表明粗网格空间相对于细网格空间可以选择很小，并且罚参数的选取只与粗网格步长和问题的正则性有关.因此罚参数不必选择很小仍能够得到最优解.最后通过数值算例验证了上述理论结果，并且数值对比可知两层罚函数方法对于求解定常不可压Stokes方程具有很好的效果.     

一类可靠性分布参数的经验Bayes决策分析

在弱平稳α-混合样本下利用概率密度函数的核估计构造了伽玛分布族参数的经验Bayes(EB)检验函数,并获得了它的渐进最优(a.o.)性.在适当的条件下证明了所提出的EB检验函数收敛速度可任意接近O(n~(-1/2)).     

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.     

随机删失下伽玛分布族参数的经验Bayes双边检验

利用经验贝叶斯方法研究了伽玛分布参数的双边检验问题,构造了一个在历史样本被随机右删失的条件下参数的经验Bayes检验函数,在适当的条件下证明了所提出的经验Bayes检验函数的渐近最优性,并获得了它的收敛速度可任意接近O(n~(-1/2)).     

Strong convergence rate of principle of averaging for jump-diffusion processes

We study jump-diffusion processes with two well-separated time scales. It is proved that the rate of strong convergence to the averaged effective dynamics is of order O(ɛ ^1/2), where ɛ ≪ 1 is the parameter measuring the disparity of the time scales in the system. The convergence rate is shown to be optimal through examples. The result sheds light on the designing of efficient numerical methods for multiscale stochastic dynamics.     

Numerical Quadratures for Hadamard Hypersingular Integrals

In this paper,we develop Gaussian quadrature formulas for the Hadamard fi- nite part integrals.In our formulas,the classical orthogonal polynomials such as Legendre and Chebyshev polynomials are used to approximate the density function f(x)so that the Gaussian quadrature formulas have degree n-1.The error estimates of the formulas are obtained.It is found from the numerical examples that the convergence rate and the accu- racy of the approximation results are satisfactory.Moreover,the rate and the accuracy can be improved by choosing appropriate weight functions.     

N-A样本下单边截断型分布族位置参数函数E·B估计的收敛速度

§ 1.　IntroductionandtheMainResult　　Asfarasweknown ,theissuesofEmpirialBayes(E·B)statisticsmainlyincludedE·Bes timationsandE·Btests ,whichwereconsideredandstudiedformerlyunderi.i.d .samples.However,insteadofgetingi.i.d .samplesinsomefieldssuchasreliabletheory ,penetratedtheoryandsomemultivariateanalysis,etc .,weoftengetsomeassociatedsamplessuchasposi tivelyassociationandnegativelyassociation .Recently ,professorWEI [3],XU [4],andLING [5 ]consideredrespectivelytheissuesofE·Btesta…     

On the Superlinear Convergence Order of the Logarithmic Barrier Algorithm

Since the pioneering work of Karmarkar, much interest was directed to penalty algorithms, in particular to the log barrier algorithm. We analyze in this paper the asymptotic convergence rate of a barrier algorithm when applied to non-linear programs. More specifically, we consider a variant of the SUMT method, in which so called extrapolation predictor steps allowing reducing the penalty parameter r_k +1}k are followed by some Newton correction steps. While obviously related to predictor-corrector interior point methods, the spirit differs since our point of view is biased toward nonlinear barrier algorithms; we contrast in details both points of view. In our context, we identify an asymptotically optimal strategy for reducing the penalty parameter r and show that if r_k+1=r _k with < 8/5, then asymptotically only 2 Newton corrections are required, and this strategy achieves the best overall average superlinear convergence order (1.1696). Therefore, our main result is to characterize the best possible convergence order for SUMT type methods.     

Frequency polygons for continuous random fields

We study the frequency polygon investigated by Scott (J Am Stat Assoc 80: 348–354, 1985) as a nonparametric density estimate for a continuous and stationary real random field \({\left( X_{\mathbf{t}},\mathbf{t}\in\mathbb{R}^{N}\right)}\). We establish the asymptotic expressions for the integrated pointwise squared bias and the integrated pointwise squared variance of the estimate when the field is observed over a rectangular domain of \({\mathbb{R}^{N}}\). Under mild mixing conditions, we show that the estimate achieves the same rate of convergence to zero of the integrated mean squared error as kernel estimators and it can also attain the optimal uniform strong rate of convergence \({\left(\widehat{\mathbf{T}}^{-1} \log \widehat{\mathbf{T}}\right)^{1/3}}\) for appropriate choices of the bin widths.     

On some methods for entropy maximization and matrix scaling

We describe and survey in this paper iterative algorithms for solving the discrete maximum entropy problem with linear equality constraints. This problem has applications e.g. in image reconstruction from projections, transportation planning, and matrix scaling. In particular we study local convergence and asymptotic rate of convergence as a function of the iteration parameter. For the trip distribution problem in transportation planning and the equivalent problem of scaling a positive matrix to achieve a priori given row and column sums, it is shown how the iteration parameters can be chosen in an optimal way. We also consider the related problem of finding a matrix X, diagonally similar to a given matrix, such that corresponding row and column norms in X are all equal. Reports of some numerical tests are given.     

On the convergence rate of a parallel nonoverlapping domain decomposition method

In recent years,a nonoverlapping domain decomposition iterative procedure,which is based on using Robin-type boundary conditions as information transmission conditions on the subdomain interfaces,has been developed and analyzed.It is known that the convergence rate of this method is 1-O(h),where h is mesh size.In this paper,the convergence rate is improved to be 1-O(h1/2 H-1/2)sometime by choosing suitable parameter,where H is the subdomain size.Counter examples are constructed to show that our convergence estimates are sharp,which means that the convergence rate cannot be better than 1-O(h1/2H-1/2)in a certain case no matter how parameter is chosen.