Bayesian Methods for Wavelet Series in Single-Index Models

Single-index models have found applications in econometrics and biometrics, where multidimensional regression models are often encountered. This article proposes a nonparametric estimation approach that combines wavelet methods for nonequispaced designs with Bayesian models. We consider a wavelet series expansion of the unknown regression function and set prior distributions for the wavelet coefficients and the other model parameters. To ensure model identifiability, the direction parameter is represented via its polar coordinates. We employ ad hoc hierarchical mixture priors that perform shrinkage on wavelet coefficients and use Markov chain Monte Carlo methods for a posteriori inference. We investigate an independence-type Metropolis-Hastings algorithm to produce samples for the direction parameter. Our method leads to simultaneous estimates of the link function and of the index parameters. We present results on both simulated and real data, where we look at comparisons with other methods.     

Stochastic expansions in an overcomplete wavelet dictionary

We consider random functions defined in terms of members of an overcomplete wavelet dictionary. The function is modelled as a sum of wavelet components at arbitrary positions and scales where the locations of the wavelet components and the magnitudes of their coefficients are chosen with respect to a marked Poisson process model. The relationships between the parameters of the model and the parameters of those Besov spaces within which realizations will fall are investigated. The models allow functions with specified regularity properties to be generated. They can potentially be used as priors in a Bayesian approach to curve estimation, extending current standard wavelet methods to be free from the dyadic positions and scales of the basis functions. Received: 21 September 1998 / Revised version: 20 August 1999 / Published online: 30 March 2000     

Cross-validated wavelet shrinkage

Cross-validation has been successfully used in various areas of statistics. However, it has not been used much in wavelet shrinkage estimation because fast wavelet methods cannot be applied to deleted data. In this paper, we show this problem can be avoided by using a fast imputation of data. This allows level-dependent cross- validation which is attractive to data with different sparseness. The proposed methods can be easily extended to higher dimensional problem such as image. Results from simulation and examples demonstrate the promising empirical properties of the procedure. In particular, the methods proposed in this work provide outstanding results for non-Gaussian noises because cross-validation is not based on normality assumptions.     

Thresholding methods to estimate copula density

This paper deals with the problem of multivariate copula density estimation. Using wavelet methods we provide two shrinkage procedures based on thresholding rules for which knowledge of the regularity of the copula density to be estimated is not necessary. These methods, said to be adaptive, have proved to be very effective when adopting the minimax and the maxiset approaches. Moreover we show that these procedures can be discriminated in the maxiset sense. We provide an estimation algorithm and evaluate its properties using simulation. Finally, we propose a real life application for financial data.     

Wavelet-Based Estimation of Anisotropic Spatiotemporal Long-Range Dependence

In this article, the estimation of spatiotemporal long-range dependence is formulated in the spectral wavelet domain. Sample information is provided by functional spectral data. Their high local singularity at the origin is captured by the wavelet transform. Weak consistency of the spectral wavelet estimators proposed is derived. Two functional estimation algorithms are implemented. A simulation study is developed to illustrate the efficiency of the computational methods derived. An approximation to the empirical convergence rate of the spectral wavelet periodogram is computed in some simulated examples.     

Adaptive and Optimal Point-Wise Estimations for Densities in GARCH-Type Model by Wavelets

This paper considers adaptive point-wise estimations of density functions in GARCH-type model under the local Holder condition by wavelet methods.A point-wise lower bound estimation of that model is first investigated;then we provide a linear wavelet estimate to obtain the optimal convergence rate,which means that the convergence rate coincides with the lower bound.The non-linear wavelet estimator is introduced for adaptivity,although it is nearly-optimal.However,the non-linear wavelet one depends on an upper bound of the smoothness index of unknown functions,we finally discuss a data driven version without any assumptions on the estimated functions.     

On the stable sampling rate for binary measurements and wavelet reconstruction

This paper is concerned with the problem of reconstructing an infinite-dimensional signal from a limited number of linear measurements. In particular, we show that for binary measurements (modelled with Walsh functions and Hadamard matrices) and wavelet reconstruction the stable sampling rate is linear. This implies that binary measurements are as efficient as Fourier samples when using wavelets as the reconstruction space. Powerful techniques for reconstructions include generalized sampling and its compressed versions, as well as recent methods based on data assimilation. Common to these methods is that the reconstruction quality depends highly on the subspace angle between the sampling and the reconstruction space, which is dictated by the stable sampling rate. As a result of the theory provided in this paper, these methods can now easily use binary measurements and wavelet reconstruction bases.     

On the Use of Particle Markov Chain Monte Carlo in Parameter Estimation of Space-Time Interacting Discs

A space-time random set is defined and methods of its parameters estimation are investigated. The evolution in discrete time is described by a state-space model. The observed output is a planar union of interacting discs given by a probability density with respect to a reference Poisson process of discs. The state vector is to be estimated together with auxiliary parameters of transitions caused by a random walk. Three methods of parameters estimation are involved, first of which is the maximum likelihood estimation (MLE) for individual outputs at fixed times. In the space-time model the state vector can be estimated by the particle filter (PF), where MLE serves to the estimation of auxiliary parameters. In the present paper the aim is to compare MLE and PF with particle Markov chain Monte Carlo (PMCMC). From the group of PMCMC methods we use specially the particle marginal Metropolis-Hastings (PMMH) algorithm which updates simultaneously the state vector and the auxiliary parameters. A simulation study is presented in which all estimators are compared by means of the integrated mean square error. New data are then simulated repeatedly from the model with parameters estimated by PMMH and the fit with the original model is quantified by means of the spherical contact distribution function.     

Robust Depth-Weighted Wavelet for Nonparametric Regression Models

In the nonparametric regression models, the original regression estimators including kernel estimator, Fourier series estimator and wavelet estimator are always constructed by the weighted sum of data, and the weights depend only on the distance between the design points and estimation points. As a result these estimators are not robust to the perturbations in data. In order to avoid this problem, a new nonparametric regression model, called the depth-weighted regression model, is introduced and then the depth-weighted wavelet estimation is defined. The new estimation is robust to the perturbations in data, which attains very high breakdown value close to 1/2. On the other hand, some asymptotic behaviours such as asymptotic normality are obtained. Some simulations illustrate that the proposed wavelet estimator is more robust than the original wavelet estimator and, as a price to pay for the robustness, the new method is slightly less efficient than the original method.     

�Dz����ع�ģ�;�ֵ�뷽��˫�ر��Ĺ���

This paper studies the estimation of change point in mean and variance function of a non-parametric regression model based on kernel estimation and wavelet method. First, kernel estimation of mean function is developed and it is used to estimate the position and jump size of mean change. Second, wavelet methods are applied to derive the variance estimator which is used to estimate the location and jump size of the change point in variance. The asymptotic properties of these estimators are proved. Finally, the results from a numerical simulations and comparison study show that validate the effectiveness of our method.     

A random model approach for the LASSO

The least absolute selection and shrinkage operator (LASSO) is a method of estimation for linear models similar to ridge regression. It shrinks the effect estimates, potentially shrinking some to be identically zero. The amount of shrinkage is governed by a single parameter. Using a random model formulation of the LASSO, this parameter can be specified as the ratio of dispersion parameters. These parameters are estimated using an approximation to the marginal likelihood of the observed data. The observed score equations from the approximation are biased and hence are adjusted by subtracting an empirical estimate of the expected value. After estimation, the model effects can be tested (via simulation) as the distribution of the observed data given that all model effects are zero is known. Two related simulation studies are presented that show that dispersion parameter estimation results in effect estimates that are competitive with other estimation methods (including other LASSO methods).     

经济时间序列的连续参数小波网络预测模型

本文利用连续小波变换方法,给出一种连续参数小波网络。网络参数的学习采用一种类似神经网络的后向传播学习算法的随机梯度算法。另外,提出了一种借助小波分析理论指导网络参数赋初值的方法。进一步,通过对中国进出口贸易额时间序列预测建模的研究和仿真预测,提出了用连续参数小波网络建立经济时间序列预测模型的一般步骤和方法。预测结果表明,此模型具有较好的泛化、学习能力,可以有效地在数值上逼近时间序列难以定量描述的相互关系,所以利用连续参数小波网络建立的时间序列预测模型有较高的预测精度。     

右删失时间序列AR模型参数估计

通过Kaplan-Meier估计和Nelson-Aalen估计得到了平稳时间序列被另一平稳序列右删失下.AR模型的参数估计.首先,通过与完全数据下的参数估计进行对比,说明了两种估计方法的效果.然后,根据计算机模拟的样本量以及删失率的不同,对比了两种估计的优劣,并且模拟结果表明两种估计是有效的.     

随机设计下非参数回归模型变点的小波检测与估计

A wavelet method of detection and estimation of change points in nonparametric regression models under random design is proposed. The confidence bound of our test is derived by using the test statistics based on empirical wavelet coefficients as obtained by wavelet transformation of the data which is observed with noise. Moreover, the consistence of the test is proved while the rate of convergence is given. The method turns out to be effective after being tested on simulated examples and applied to IBM stock market data.     

Stable Estimation in Dimension Reduction

We introduce stable estimation procedures for several aspects of a sufficient dimension-reduction matrix. We first propose a stable method for estimating structural dimension, which only selects the correct directions in the central subspace with no false positive selection. We then provide a Grassmann manifold sparse estimate for the central subspace. By using subsampling, we develop an ensemble method to obtain a stable nonsparse estimate for the central subspace. This ensemble idea is also used to stabilize the choice of the number of slices in sliced inverse methods. Theoretical results are established, and the efficacy of the proposed stable methods is demonstrated by simulation studies and the analysis of Hitters’ salary data. Supplementary materials for this article are available online.     

基于贝叶斯分析的中国商业银行内部欺诈分析

内部欺诈事件类型是中国商业银行最严重的操作风险类型。但由于操作风险本质特征和中国商业银行内部欺诈损失数据收集年度较短,数据匮乏,小样本数据容易导致参数结果不稳定。为了在小样本数据下进行更准确的度量,本文采用贝叶斯马尔科夫蒙特卡洛模拟方法,在损失分布法框架下,假设损失频率服从泊松-伽马分布,而损失强度服从广义帕累托-混合伽马分布,分析后验分布的形式,获得中国商业银行不同业务线的内部欺诈损失频率和损失强度的后验分布估计,并进行蒙特卡罗模拟获得不同业务线内部欺诈的风险联合分布。结果表明,拟合结果很好,与传统极值分析法相比,基于利用贝叶斯的分析获得的后验分布可以作为未来的先验分布,有利于在较小样本下获得较真实的参数估计,本方法有助于银行降低监管资本要求。     

A RECURSIVE ALGORITHM AND ITS CONVERGENCE FOR PARAMETER ESTIMATION OF CONVOLUTION MODEL

In this article, the problem on the estimation of the convolution model parameters is considered. The recursive algorithm for estimating model parameters is introduced from the orthogonal procedure of the data, the convergence of this algorithm is theoretically discussed, and a sufficient condition for the convergence criterion of the orthogonal procedure is given. According to this condition, the recursive algorithm is convergent to model wavelet A- = （1, α1,..., αq）.     

A bayesian approach to binary response curve estimation

Summary The purpose of the present paper is to propose a practical procedure for the estimation of the binary response curve. The procedure is based on a model which approximates the response curve by a finely segmented piecewise constant function. To obtain a stable estimate we assume a prior distribution of the parameters of the model. The prior distribution has several parameters (hyper-parameters) which are chosen to minimize an information criterion ABIC. The procedure is applicable to data consisting of observations of a binary response variable and a single explanatory variable. The practical utility of the procedure is demonstrated by examples of applications to the dose response curve estimation, to the intensity function estimation of a point process and to the analysis of social survey data. The application of the procedure to the discriminant analysis is also briefly discussed. The Institute of statistical mathematics     

Shrinkage estimator in normal mean vector estimation based on conditional maximum likelihood estimators

Estimation of normal mean vector has broad applications such as small area estimation, estimation of nonparametric functions and estimation of wavelet coefficients. In this paper, we propose a new shrinkage estimator based on conditional maximum likelihood estimator incorporating with Stein’s risk unbiased estimator (SURE) when data have the normality. We present some theoretical work and provide numerical studies to compare with some existing methods.     

Wavelet-Based Weighted LASSO and Screening Approaches in Functional Linear Regression

One useful approach for fitting linear models with scalar outcomes and functional predictors involves transforming the functional data to wavelet domain and converting the data-fitting problem to a variable selection problem. Applying the LASSO procedure in this situation has been shown to be efficient and powerful. In this article, we explore two potential directions for improvements to this method: techniques for prescreening and methods for weighting the LASSO-type penalty. We consider several strategies for each of these directions which have never been investigated, either numerically or theoretically, in a functional linear regression context. We compare the finite-sample performance of the proposed methods through both simulations and real-data applications with both 1D signals and 2D image predictors. We also discuss asymptotic aspects. We show that applying these procedures can lead to improved estimation and prediction as well as better stability. Supplementary materials for this article are available online.