Lp( Rn)中的神经网络逼近和系统识别

本文主要研究函数的叠加对Lp(Rn)中的函数,Lp(Rn)上的非线性连续泛函及非线性连续算子的逼近.这些问题与Sigma-Pi型神经网络逼近能力有关.     

Lp( Rn)中的神经网络逼近和系统识别

本文主要研究函数的叠加对Lp(Rn)中的函数, Lp(Rn)上的非线性连续泛函及非线性连续算子的逼近. 这些问题与Sigma-Pi 型神经网络逼近能力有关.     

W_2~m(R~n)中连续泛函及连续算子的神经网络逼近

本文主要讨论定义在Sobolev空间Wm2 (Rn)中紧集上连续泛函及连续算子的逼近 .     

结构可靠性分析的支持向量机方法

针对结构可靠性分析中功能函数不能显式表达的问题,将支持向量机方法引入到结构可靠性分析中．支持向量机是一种实现了结构风险最小化原则的分类技术,它具有出色的小样本学习性能和良好的泛化性能,因此提出了两种基于支持向量机的结构可靠性分析方法．与传统的响应面法和神经网络法相比,支持向量机可靠性分析方法的显著特点是在小样本下高精度地逼近函数,并且可以避免维数灾难．算例结果也充分表明支持向量机方法可以在抽样范围内很好地逼近真实的功能函数,减少隐式功能函数分析（通常是有限元分析）的次数,具有一定的工程实用价值．     

卷积算子交换子的Lp(Rn)有界性

本文利用Foulrier变换估计和逼近恒等,建立由BMo函数和卷积算子生成的交换子的Lp(Rn)有界性结果.作为一个应用,得到了关于Fefferman型奇异积分算子交换子的Lp(Rn)有界性的一个新的结果.     

一种新的光滑支持向量机的构造及其应用

光滑支持向量机模型是一个无约束、可微的最优化模型,人们可应用快速的最优化方法求解,从而降低计算复杂性.在前人工作的基础上研究基于样条函数的光滑支持向量机,采用广义三弯矩方法构造出六次样条光滑函数,分析了其性能及与正号函数的逼近精度,实现了求解六次样条光滑支持向量机的算法,与其它光滑支持向量机进行了比较,取得了较好的结果.最后将其应用于心脏病模型诊断,实验结果显示具有较高的精确度.     

Banach空间中有限秩拟线性投影算子的逼近问题

证得:在Banach空间中,相对紧集上的恒等算子可由一列有限秩连续拟线性投影算子一致逼近.由此得到:线性算子为紧线性算子必须且仅须它可由一列有限秩连续齐性算子一致逼近.     

基于Fuzzy理论的数据挖掘算法研究(Ⅰ)

“数据挖掘”是数据处理的一个新领域.支持向量机是数据挖掘的一种新方法,该技术在很多领域得到了成功的应用.但是,支持向量机目前还存在许多局限,当支持向量机的训练集中含有模糊信息时,支持向量机将无能为力.为解决一般情况下支持向量机中含有模糊信息(模糊参数)问题,研究了模糊机会约束规划、模糊分类中的模糊特征及其表示方法,建立了模糊支持向量分类机理论,给出了模糊线性可分的模糊支持向量分类机算法.     

基于Fuzzy理论的数据挖掘算法研究Ⅱ

"数据挖掘"是数据处理的一个新领域.支持向量机是数据挖掘的一种新方法,该技术在很多领域得到了成功的应用.但是,支持向量机目前还存在许多局限,当支持向量机的训练集中含有模糊信息时,支持向量机将无能为力.为解决一般情况下支持向量机中含有模糊信息(模糊参数)问题,研究了模糊机会约束规划、模糊分类中的模糊特征及其表示方法,建立了模糊支持向量分类机理论,给出了模糊线性可分的模糊支持向量分类机算法.     

双障碍问题的逐次逼近阻尼牛顿法

1．引言及算法考虑Rn中的双障碍问题：求＊E年使其中f：Rn→Rn连续可微,c＝｛x-∈Rn|0≤x≤c｝,c∈Rn为常向量．若记c＝（c1,…,cn）T,则不准证明问题（1）等价于求解下面的非光滑方程组其中算子max,min是指分量的最大或最小．显然,由（2）式定义的函数H：Rn→Rn的第i个分量函数Hi：Rn→R为由（2）定义的函数H一般不是厂可微的,但我们可对H作如下分解：设住k｝是一单调递减且趋于O的正数序列,xk00,八k（）都是R”－＋R”的映射,其中bk（儿,W汕地分别由下式定义：其中圳的二v：二;＜人（x）一。k｝,B（x）二F：人（……     

A Method for Solving Initial Value Problems for Linear Differential Equations in Hilbert Space Based on The Cayley Transform

A first-order differential equation with a bounded operator coefficient in Hilbert space is considered. In case the coefficient is constant we establish a one-to-one correspondence between the continuous initial value problem and some discrete initial value problem. Using the Cayley transform we give explicit formulas for their solutions as well as formulas connecting corresponding continuous and discrete semigroups. On the basis of these formulas we propose a numerical algorithm for solving initial value problems with a bounded constant operator coefficient which has an exponential rate of convergence. If the operator coefficient is variable we use its piecewise constant approximation and the previously cited algorithm. Error estimates are given.     

Reconstruction of vector fields and their singularities from ray transforms

In this paper, numerical methods for reconstruction of the singular support of a vector field from its known longitudinal and (or) transverse ray transforms are proposed. Apart from a modification for the Vainberg operator, we use integral operators of angular moments and back projections as well as differential operators of tensor analysis for solving the problem. Results of numerical simulation for reconstructing discontinuous vector fields and with discontinuities in the derivatives are presented. Visualization of their singular support is shown.     

Stability results for the approximation of weakly coupled wave equations

In this Note, we consider the approximation of two coupled wave equations with internal damping. Our goal is to damp the spurious high frequency modes by introducing numerical viscosity terms in the approximation scheme. With these viscosity terms, we show the exponential or polynomial decay of the discrete scheme when the continuous problem has such a decay (since the spectrum of the spatial operator associated with the undamped system satisfies the generalized gap condition).     

基于DC规划方法的稀疏临近支持向量机

为了提高临近支持向量机（PSVM）的数值表现,在PSVM的模型中引入了$\ell_0$-范数正则项,提出了稀疏临近支持向量机模型（SPSVM）,从而提高分类器的特征选择能力。然而带有$\ell_0$-范数正则项的问题往往是NP-难问题,为了克服这一问题,采用非凸连续函数近似$\ell_0$-范数,并通过适当的DC分解将问题转化成DC规划问题进行求解,同时还讨论了算法的收敛性。数值实验结果表明不论是在仿真数据还是在实际数据中,所提出的方法是比较有效稳定的。     

Multi-kernel regularized classifiers

A family of classification algorithms generated from Tikhonov regularization schemes are considered. They involve multi-kernel spaces and general convex loss functions. Our main purpose is to provide satisfactory estimates for the excess misclassification error of these multi-kernel regularized classifiers when the loss functions achieve the zero value. The error analysis consists of two parts: regularization error and sample error. Allowing multi-kernels in the algorithm improves the regularization error and approximation error, which is one advantage of the multi-kernel setting. For a general loss function, we show how to bound the regularization error by the approximation in some weighted L^q$L^q$ spaces. For the sample error, we use a projection operator. The projection in connection with the decay of the regularization error enables us to improve convergence rates in the literature even for the one-kernel schemes and special loss functions: least-square loss and hinge loss for support vector machine soft margin classifiers. Existence of the optimization problem for the regularization scheme associated with multi-kernels is verified when the kernel functions are continuous with respect to the index set. Concrete examples, including Gaussian kernels with flexible variances and probability distributions with some noise conditions, are used to illustrate the general theory.     

Numerical Solution of Time-Dependent Problems with a Fractional-Power Elliptic Operator

A time-dependent problem in a bounded domain for a fractional diffusion equation is considered. The first-order evolution equation involves a fractional-power second-order elliptic operator with Robin boundary conditions. A finite-element spatial approximation with an additive approximation of the operator of the problem is used. The time approximation is based on a vector scheme. The transition to a new time level is ensured by solving a sequence of standard elliptic boundary value problems. Numerical results obtained for a two-dimensional model problem are presented.     

Spectral computations: from operators to matrices

In this work we will address the problem of finding spectral values and bases for the corresponding spectral subspaces for a bounded operator on a Banach space. We will make a bridge between the spectral problem for the continuous operator and the computation of the eigenpairs of a matrix. This approximate problem results first from an approximation by a sequence of continuous operators with finite rank, followed by the reduction to a spectral problem for an operator whose domain as well as range are finite dimensional. Some discussion on defect correction procedures will be also presented. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Vector scheduling with rejection on a single machine

In this paper, we study a vector scheduling problem with rejection on a single machine, in which each job is characterized by a d-dimension vector and a penalty, in the sense that, jobs can be either rejected by paying a certain penalty or assigned to the machine. The objective is to minimize the sum of the maximum load over all dimensions of the total vector of all accepted jobs, and the total penalty of rejected jobs. We prove that the problem is NP-hard and design two approximation algorithms running in polynomial time. When d is a fixed constant, we present a fully polynomial time approximation scheme.     

Decent flow methods for inverse Sturm–Liouville problem

In this paper, we propose three numerical methods for the inverse Sturm–Liouville operator in impedance form. We use a finite difference method to discretize the Sturm–Liouville operator and expand the impedance function with some basis functions. The correction technique is discussed. By solving an un-weighted least squares problem, we find an approximation to the impedance function. Numerical experiments are presented to show the accuracy and stability of the numerical methods.     

Sufficient conditions of non‐uniqueness for the Coulomb friction problem

We consider the Signorini problem with Coulomb friction in elasticity. Sufficient conditions of non‐uniqueness are obtained for the continuous model. These conditions are linked to the existence of real eigenvalues of an operator in a Hilbert space. We prove that, under appropriate conditions, real eigenvalues exist for a non‐local Coulomb friction model. Finite element approximation of the eigenvalue problem is considered and numerical experiments are performed. Copyright © 2003 John Wiley & Sons, Ltd.