部分线性分位回归模型估计的MM算法

近年来,关于部分线性分位回归模型的估计方法的研究得到了较多的关注.但由于目标函数的非光滑性,估计程序的实现是比较具有挑战性的.文章将采用MM(Majorization Minimization)算法计算部分线性分位数回归模型的估计.其基本原理是首先找到目标函数的优化函数,然后借助优化函数的最小化过程,逐步迭代至目标函数的解.数值模拟和实证研究表明该算法具有较好的稳定性和较强的数值计算能力.     

基于固定效应的纵向数据分位点回归模型的参数估计及CDM和MSOM的等价性

研究了基于固定效应的纵向数据模分位点回归模型的参数估计及统计诊断问题.首先给出了参数估计的MM迭代算法,然后讨论了统计诊断中数据删除模型(CDM)和均值移模型(MSOM)的等价性问题,最后利用消炎镇痛药数据说明了方法的应用.     

基于偏正态数据下联合位置与尺度混合专家回归模型的参数估计北大核心CSCD

混合专家回归模型广泛应用于异质总体数据的分类,聚类及回归分析中.研究基于偏正态数据,提出了联合位置与尺度混合专家回归模型,该模型同时对位置,尺度和混合比例参数建模,应用MM算法和EM算法研究了该模型参数的极大似然估计.通过随机模拟和实例分析说明了该模型和方法的有效性与实用性.     

在线Logistic回归

针对连续数据流分类问题,基于在线学习理论,提出一种在线logistic回归算法.研究带有正则项的在线logistic回归,提出了在线logistic-l2回归模型,并给出了理论界估计.最终实验结果表明,随着在线迭代次数的增加,提出的模型与算法能够达到离线预测的分类结果.本文工作为处理海量流数据分类问题提供了一种新的有效方法.     

基于偏正态数据下联合位置与尺度混合专家回归模型的参数估计

混合专家回归模型广泛应用于异质总体数据的分类,聚类及回归分析中.研究基于偏正态数据,提出了联合位置与尺度混合专家回归模型,该模型同时对位置,尺度和混合比例参数建模,应用MM算法和EM算法研究了该模型参数的极大似然估计.通过随机模拟和实例分析说明了该模型和方法的有效性与实用性.     

基于字典学习和TV的能谱CT重建算法

能谱CT可以将较宽的能谱数据划分为几个单独的窄谱数据,从而同时获得多个能量通道下的投影.但由于窄谱通道内接收到的光子数较少,投影通常包含较大的噪声.针对这一问题,基于压缩感知理论提出了一种基于字典学习和全变分TV(total-variation)的迭代重建算法用于能谱CT重建,应用交替最小化方法优化相关目标函数,并采用Split-Bregman算法求解.同时,采用有序子集方法加速迭代收敛过程,提高运算速率.为了验证和评估所提出的方法,使用简单模型和实际临床小鼠模型进行了仿真实验,实验结果表明,所提出的算法有较好的去噪及细节保存能力.     

基于新的Hessian近似矩阵的稀疏重构算法

一般来说,基于二次近似模型的优化算法具有良好的数值表现.然而,当基于二次近似模型的优化算法求解大规模优化问题时,若使用稠密矩阵近似目标函数在迭代点的Hessian矩阵,需要花费大量的计算成本和存储成本,因此设计Hessian矩阵合适的标量近似矩阵特别重要.对于正则化模型,利用最近三次迭代的信息,设计粗糙的标量矩阵,使用拟牛顿公式进行更新,结合近似最优梯度法的思想和梯度法的延迟策略,构造Hessian矩阵新的含有更多二阶信息的标量近似矩阵.结合非单调线搜索,提出基于新的Hessian近似矩阵的稀疏重构算法,并进行收敛性分析.实验结果表明,与经典稀疏重构算法算法相比,基于新的Hessian近似矩阵的稀疏重构算法在重构效果相似的情况下能较大地减少迭代次数和较快地重构信号.     

基于指数Laplace损失函数的回归估计鲁棒超限学习机

实际问题的数据集通常受到各种噪声的影响,超限学习机(extreme learning machine, ELM)对这类数据集进行学习时,表现出预测精度低、预测结果波动大.为了克服该缺陷,采用了能够削弱噪声影响的指数Laplace损失函数.该损失函数是建立在Gauss核函数基础上,具有可微、非凸、有界且能够趋近于Laplace函数的特点.将其引入到超限学习机中,提出了鲁棒超限学习机回归估计(exponential Laplace loss function based robust ELM for regression, ELRELM)模型.利用迭代重赋权算法求解模型的优化问题.在每次迭代中,噪声样本点被赋予较小的权值,能够有效地提高预测精度.真实数据集实验验证了所提出的模型相比较于对比算法具有更优的学习性能和鲁棒性.     

一个结构型凸优化问题的分裂算法

考虑带线性约束的三块变量的凸优化模型,目标函数是可分的三个函数和.给出了一个新的分裂算法.首先,对每个块变量解极小化增广拉格朗日函数.然后,通过一个校正步得到新的迭代点.证明了新算法的整体收敛性和O(1/t)的收敛阶.     

比例风险模型下参数估计的两种MM算法的一些应用

对于带有删失机制的生存数据的研究,比例风险模型是应用最为广泛的统计模型之一。实际中,为得到其参数的极大似然估计需要采用数值方法计算得分方程的解。MinorizationMaximization算法(以下简称"MM算法")将求解复杂的目标函数的极值问题转化为求解简单的代理函数的极值问题。本文主要探讨,在比例风险模型下通过两种不同的思想为偏似然函数构造代理函数,从而得到的两种MM算法。通过数值模拟和实际数据分析实现这两种MM算法在比例风险模型下的一些应用。     

A REPRESENTATION FORMULA RELATED TO SCHR(O)DINGER OPERATORS

Schr(o)dinger operator is a central subject in the mathematical study of quantum mechanics.Consider the Schrodinger operator H = -△ V on R, where △ = d2/dx2 and the potential function V is real valued. In Fourier analysis, it is well-known that a square integrable function admits an expansion with exponentials as eigenfunctions of -△. A natural conjecture is that an L2 function admits a similar expansion in terms of "eigenfunctions" of H, a perturbation of the Laplacian (see [7], Ch. Ⅺ and the notes), under certain condition on V.     

On a class of self-similar processes with stationary increments in higher order Wiener chaoses

We study a class of self-similar processes with stationary increments belonging to higher order Wiener chaoses which are similar to Hermite processes. We obtain an almost sure wavelet-like expansion of these processes. This allows us to compute the pointwise and local Hölder regularity of sample paths and to analyse their behaviour at infinity. We also provide some results on the Hausdorff dimension of the range and graphs of multidimensional anisotropic self-similar processes with stationary increments defined by multiple Wiener–Itô integrals.     

Qualitative Analysis of Bifurcating Solutions in the Lengyel-Epstein Model

One of the most fundamental problems in theoretical biology is to explain the mechanisms by which patterns and forms are created in the＇living world. In his seminal paper ＂The Chemical Basis of Morphogenesis＂, Turing showed that a system of coupled reaction-diffusion equations can be used to describe patterns and forms in biological systems. However, the first experimental evidence to the Turing patterns was observed by De Kepper and her associates（1990） on the CIMA reaction in an open unstirred reactor, almost 40 years after Turing＇s prediction. Lengyel and Epstein characterized this famous experiment using a system of reaction-diffusion equations. The Lengyel-Epstein model is in the form as follows     

(0,1 ;0)-INTERPOLATION ON INFINITE INTERVAL (-∞, +∞)

In this paper, we study the explicit representation and convergence of (0, 1; 0)-interpolation on infinite interval, which means to determine a polynomial of degree ≤ 3n - 2 when the function values are prescribed at two set of points namely the zeros of Hn(x) and H′n(x) and the first derivatives at the zeros of H′n(x).     

On a class of Riemann surfaces

It is considered the class of Riemann surfaces with dimT1 = 0, where T1 is a subclass of exact harmonic forms which is one of the factors in the orthogonal decomposition of the spaceΩH of harmonic forms of the surface, namely The surfaces in the class OHD and the class of planar surfaces satisfy dimT1 = 0. A.Pfluger posed the question whether there might exist other surfaces outside those two classes. Here it is shown that in the case of finite genus g, we should look for a surface S with dimT1 = 0 among the surfaces of the form Sg\K , where Sg is a closed surface of genus g and K a compact set of positive harmonic measure with perfect components and very irregular boundary.     

CONTENTS OF VOLUME 30

正~~     

Instructions for contributions

正Applied Mathematics-A Journal of Chinese Universities,Series B(Appl.Math.J.Chinese Univ.,Ser.B)is a comprehensive applied mathematics journal jointly sponsored by Zhejiang University,China Society for Industrial and Applied Mathematics,and Springer-Verlag.It is a quarterly journal with     

Journal of Mathematical Research with Applications Guide for Authors

正Journal overview:Journal of Mathematical Research with Applications(JMRA),formerly Journal of Mathematical Research and Exposition(JMRE)created in 1981,one of the transactions of China Society for Industrial and Applied Mathematics,is a home for original research papers of the highest quality in all areas of mathematics with applications.The target audience comprises:pure and applied mathematicians,graduate students in broad fields of sciences and technology,scientists and engineers interested in mathematics.     

Staircase transportation problems with superadditive rewards and cumulative capacities

A cumulative-capacitated transportation problem is studied. The supply nodes and demand nodes are each chains. Shipments from a supply node to a demand node are possible only if the pair lies in a sublattice, or equivalently, in a staircase disjoint union of rectangles, of the product of the two chains. There are (lattice) superadditive upper bounds on the cumulative flows in all leading subrectangles of each rectangle. It is shown that there is a greatest cumulative flow formed by the natural generalization of the South-West Corner Rule that respects cumulative-flow capacities; it has maximum reward when the rewards are (lattice) superadditive; it is integer if the supplies, demands and capacities are integer; and it can be calculated myopically in linear time. The result is specialized to earlier work of Hoeffding (1940), Fréchet (1951), Lorentz (1953), Hoffman (1963) and Barnes and Hoffman (1985). Applications are given to extreme constrained bivariate distributions, optimal distribution with limited one-way product substitution and, generalizing results of Derman and Klein (1958), optimal sales with age-dependent rewards and capacities.To our friend, Philip Wolfe, with admiration and affection, on the occasion of his 65th birthday.Research was supported respectively by the IBM T.J. Watson and IBM Almaden Research Centers and is a minor revision of the IBM Research Report [6].