判别分析与Logistic回归组合分类

本文讨论判别分析与Logistic回归组合方法的二分类问题.模拟结果显示,基于判别分析的Logistic回归的第一组合法的回判正确率往往高于判别分析分类和Logistic回归分类.第二组合法是基于Logistic回归的判别分析分类,回判正确率接近于Logistic回归的分类结果.很多情况下,这两种方法的回判正确率都高于判别分析。     

线性ε-支持向量回归机的数据扰动分析

支持向量回归机是解决回归问题的一个重要方法.在实际问题中由于测量及计算误差的存在,我们得到的数据往往只是真值的某种近似,带有一定的舍入误差,因此有必要研究支持向量回归机的数据扰动问题.考虑到线性回归问题在实际生活中有广泛的应用价值,把线性ε-支持向量回归机作为研究对象.由于最终关心的是它的原始问题的解,所以我们研究给定的训练集中输入数据发生微小地扰动后,原始问题的解的变化情况.在一定的条件下给出了解对扰动数据偏导数的表达式,建立了线性ε-支持向量回归机的原始问题的灵敏度分析定理.文中还进一步分析了建立该灵敏度分析定理所需要的条件,给出了条件减弱后的结果.文章最后还通过一些简单的数值试验验证了定理的准确性.     

借助定义，妙破问题

<正>定义是相关知识的理论基础和精神灵魂,借助定义,回归本质,往往是破解问题的一个基本切入点.圆锥曲线中的抛物线,其定义很好地反映了曲线的本质特征,揭示了曲线存在的几何性质与规律.回归抛物线定义,应用抛物线定义,实现抛物线上的点到焦点的距离与该点到准线距离之间的合理转化,即实现“两点距离”和“点线距离”之间的合理转化,是破解抛物线问题中非常常用的一个技巧方法.1 线段长度的计算     

复合分位数下门限自回归模型的变点估计

金融领域中的突发事件是变点问题的一种体现,往往由于其随机性和发生前信息量不足等因素造成突发事件难以识别和预测.金融市场常常表现出非线性和异质性等特征,门限分位数自回归模型作为金融领域变点问题研究的重要模型,逐渐在经济和统计学界获得更多的关注.本文结合不同的分位数对门限分位数自回归模型中的交点估计问题提出两种新的估计方法...     

谈谈“线性回归”的教学

新教材选修Ⅱ概率与统计第六节线性回归,介绍了变量间的相关关系中最简单的一种形式:一元线性相关关系.教学要求学生通过学习,了解线性回归的基本思想和方法.但在教学实践中,许多教师往往回避线性方程的意义和形成过程,只简单扼要地引出公式和结论,把教学的重点     

中位数回归的贝叶斯变量选择方法

当数据呈现厚尾特征或含有异常值时,基于惩罚最小二乘或似然函数的传统变量选择方法往往表现不佳.本文基于中位数回归和贝叶斯推断方法,研究线性模型的贝叶斯变量选择问题.通过选取回归系数的Spike and Slab先验,利用贝叶斯模型选择理论提出了中位数回归的贝叶斯估计方法,并提出了有效的后验Gibbs抽样程序.大量数值模拟和波士顿房价数据分析充分说明了所提方法的有效性.     

删失混合效应模型的分位回归及变量选择

纵向数据常常用正态混合效应模型进行分析.然而,违背正态性的假定往往会导致无效的推断.与传统的均值回归相比较,分位回归可以给出响应变量条件分布的完整刻画,对于非正态误差分布也可以给稳健的估计结果.本文主要考虑右删失响应下纵向混合效应模型的分位回归估计和变量选择问题.首先,逆删失概率加权方法被用来得到模型的参数估计.其次,结合逆删失概率加权和LASSO惩罚变量选择方法考虑了模型的变量选择问题.蒙特卡洛模拟显示所提方法要比直接删除删失数据的估计方法更具优势.最后,分析了一组艾滋病数据集来展示所提方法的实际应用效果.     

纵向删失数据下线性模型的相合估计

医药临床试验,生存分析,可靠性统计等研究领域,由于考虑到时间和费用问题,研究往往有一定期限.因为研究到期的被迫结束或者某些病人中途退出试验,最后得到的试验结果往往是删失数据.对于删失数据,采用无偏转换的方法处理,方法的最大优点是得到的估计量为显式解.首先讨论了在纵向右删失数据下线性回归模型回归系数估计的均方相合性,并且把结论推广到了污染线性模型,得到了污染系数、回归系数的强相合估计.     

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.     

估计极端行为模型:分位数回归方法及其实现与应用

在许多社会和管理研究中,研究者通常很感兴趣不同于期望或平均的极端行为的理论解释。这些特殊个案所包含的信息往往是研究的创新点和解决某些问题的突破口,但传统的最小平方法与最小一乘法并不适宜于这类研究问题的解决。本文讨论一种估计极端行为的理想模型:分位数回归。本文在对分位数回归的国内外研究现状进行综述后,介绍了分位数回归的模型和实现方法,并将它与最小平方法、最小一乘法进行了比较。最后探讨它在我国管理研究领域的应用方式和有关条件。     

Identification of distributed systems and the theory of the regularization

The identification problems, i.e., the problems of finding unknown parameters in distributed systems from the observations are very important in modern control theory. The solutions of these identification problems can be obtained by solving the equations of the first kind. However, the solutions are often unstable. In other words, they are not continuously dependent on the data. The regularization or Tihonov's regularization is known as one of the stabilizing algorithms to solve these non well-posed problems. In this paper is studied the regularization method for identification of distributed systems. Several approximation theorems are proved to solve the equations of the first kind. Then, identification problems are reduced to the minimization of quadratic cost functionals by virtue of these theorems. On the other hand, it is known that the statistical methods for identification such as the maximum likelihood lead to the minimization problems of certain quadratic functionals. Comparing these quadratic cost functionals, the relations between the regularization and the statistical methods are discussed. Further, numerical examples are given to show the effectiveness of this method.     

On the completeness of eigenfunctions of the Frankl problem with the oddness condition

Modified Frankl problems with the oddness condition are considered for the Lavrent’ev-Bitsadze equation. The eigenvalues and eigenfunctions of these problems are found. The completeness of these eigenfunctions in the elliptic part of the domain is proved.     

由三个特殊次序向量对构造三对角对称矩阵

讨论利用给定的三个特殊次序向量对构造不可约三对角矩阵、Jacobi矩阵和负Jacobi矩阵的反问题.在求解方法中,将已知的一些关系式等价地转化为线性方程组,利用线性方程组有解的条件,得到了所研究问题有惟一解的充要条件,并给出了数值算法和例子.     

Conditions Characterizing Minima of the Difference of Functions

 Optimization problems involving differences of functions arouse interest as generalizations of so-called d.c. problems, i.e. problems involving the difference of two convex functions. The class of d.c. functions is very rich, so d.c. problems are rather general optimization problems. Several global optimality conditions for these d.c. problems have been proposed in the optimization literature. We provide a survey of these conditions and try to detect their common basis. This enables us to give generalizations of the conditions to situations when the objective function is no longer a difference of convex functions, but the difference of two functions which are representable as the upper envelope of an arbitrary family of functions. (Received 6 February 2001; in revised form 11 October 2001)     

Toward the Routine Analysis of Diverse Data Types

This article describes a variety of data analysis problems. The types of data across these problems included free text, parallel text, an image collection, remote sensing imagery, and network packets. A strategy for approaching the analysis of these diverse types of data is described. A key part of the challenge is mapping the analytic results back into the original domain and data setting. Additionally, a common computational bottleneck encountered in each of these problems is diagnosed as analysis tools and algorithms with unbounded memory characteristics. This experience and the analysis suggest a research and development path that could greatly extend the scale of problems that can be addressed with routine data analysis tools. In particular, there are opportunities associated with developing theory and functioning algorithms with favorable memory-usage characteristics, and there are opportunities associated with developing methods and theory for describing the outcomes of analyses for the various types of data.     

Solving a suite of NIST benchmark problems for adaptive FEM with the Hermes library

Recently, a new suite of twelve benchmark problems for adaptive finite element methods (FEM) was published at the US National Institute for Standards and Technology (NIST). These benchmark problems come with exact solutions, and they exhibit all typical difficulties associated with elliptic problems including singularities, steep internal layers, anisotropy, and oscillations. In this paper we solve these benchmark problems using the open source library Hermes (http://hpfem.org). All these results are reproducible—they are part of the Git repository of the open source Hermes project, and the reader can experiment with them by himself/herself. Instructions for this are provided. We hope that authors of other adaptive FEM codes will make their results for these test problems available in a reproducible fashion as well.     

关于两参数常微分方程非线性边值问题的奇异摄动

本文应用改进的多重尺度法及特殊的对多参数问题的讨论方法，考察含两个小参数的常微分方程的非线性边值问题，证明了解的存在性，作出了渐近解，并应用较简单的方法估计了余项。     

On the convergence of the generalized Weiszfeld algorithm

In this paper we consider Weber-like location problems. The objective function is a sum of terms, each a function of the Euclidean distance from a demand point. We prove that a Weiszfeld-like iterative procedure for the solution of such problems converges to a local minimum (or a saddle point) when three conditions are met. Many location problems can be solved by the generalized Weiszfeld algorithm. There are many problem instances for which convergence is observed empirically. The proof in this paper shows that many of these algorithms indeed converge.     

Functionals on the spaces of convex bodies

In geometry, there are several challenging problems studying numbers associated to convex bodies. For example, the packing density problem, the kissing number problem, the covering density problem, the packing-covering constant problem, Hadwiger's covering conjecture and Borsuk's partition conjecture. They are fundamental and fascinating problems about the same objects. However, up to now, both the methodology and the technique applied to them are essentially different. Therefore, a common foundation for them has been much expected. By treating problems of these types as functionals defined on the spaces of n-dimensional convex bodies, this paper tries to create such a foundation. In particular, supderivatives for these functionals will be studied.     

Stability estimates for the solutions to inverse extremal problems for the Helmholtz equation

The inverse problems are under study for the Helmholtz equation describing acoustic scattering at a three-dimensional inclusion. Some optimization method reduces these problems to the inverse extremum problems with variable refraction index and boundary source density as controls. We prove that these problems are solvable and derive the optimality systems that describe necessary optimality conditions. Analysis of the optimality systems leads us to some sufficient conditions on the input data ensuring the uniqueness and stability of optimal solutions.