Tolerance regions in multivariate linear regression

The construction of tolerance regions is investigated for a multivariate linear regression model under the multivariate normality assumption. In the context of such a model, a tolerance region is a region that will contain, with a certain confidence, at least a specified proportion of the population distribution, for a fixed value of the independent variable in the regression model. The necessary framework is developed for the Monte Carlo estimation of the tolerance factor. Some approximations are developed for the tolerance factor, and the accuracy of the approximations is numerically investigated. The approximations provide tolerance factors that are quite easy to compute, and the numerical results indicate the situations where the approximations are satisfactory. The computations are illustrated using an example.     

Approximations of Variances and Covariances for Order Statistics from the Standard Extreme Value Distribution

We consider simple approximations of variances and covariances for order statistics from the standard extreme value distribution. Exact values and simulation results of the variances and covariances for certain sample sizes are used to determine the validity of the suggested approximations.     

Asymptotic tests and monte carlo studies associated with the multiplicative interaction model

Several authors have proposed approximations to percentage points required for testing certain hypotheses associated with the multiplicative interaction model. Alternative approximations based on the asymptotic joint distribution of the characteristic roots of a noncentral Wishart matrix are proposed in this paper. The type I error rates of the resulting tests and the existing procedures are then compared using Monte Carlo methods.     

Percentiles of sums of heavy-tailed random variables: beyond the single-loss approximation

A perturbative approach is used to derive approximations of arbitrary order to estimate high percentiles of sums of positive independent random variables that exhibit heavy tails. Closed-form expressions for the successive approximations are obtained both when the number of terms in the sum is deterministic and when it is random. The zeroth order approximation is the percentile of the maximum term in the sum. Higher orders in the perturbative series involve the right-truncated moments of the individual random variables that appear in the sum. These censored moments are always finite. As a result, and in contrast to previous approximations proposed in the literature, the perturbative series has the same form regardless of whether these random variables have a finite mean or not. For high percentiles, and specially for heavier tails, the quality of the estimate improves as more terms are included in the series, up to a certain order. Beyond that order the convergence of the series deteriorates. Nevertheless, the approximations obtained by truncating the perturbative series at intermediate orders are remarkably accurate for a variety of distributions in a wide range of parameters.     

Smoothing spline Gaussian regression: more scalable computation via efficient approximation

Summary.  Smoothing splines via the penalized least squares method provide versatile and effective nonparametric models for regression with Gaussian responses. The computation of smoothing splines is generally of the order O ( n ³), n being the sample size, which severely limits its practical applicability. We study more scalable computation of smoothing spline regression via certain low dimensional approximations that are asymptotically as efficient. A simple algorithm is presented and the Bayes model that is associated with the approximations is derived, with the latter guiding the porting of Bayesian confidence intervals. The practical choice of the dimension of the approximating space is determined through simulation studies, and empirical comparisons of the approximations with the exact solution are presented. Also evaluated is a simple modification of the generalized cross-validation method for smoothing parameter selection, which to a large extent fixes the occasional undersmoothing problem that is suffered by generalized cross-validation.     

Sparse Approximations of Fractional Matérn Fields

We consider fast lattice approximation methods for a solution of a certain stochastic non‐local pseudodifferential operator equation. This equation defines a Matérn class random field. We approximate the pseudodifferential operator with truncated Taylor expansion, spectral domain error functional minimization and rounding approximations. This allows us to construct Gaussian Markov random field approximations. We construct lattice approximations with finite‐difference methods. We show that the solutions can be constructed with overdetermined systems of stochastic matrix equations with sparse matrices, and we solve the system of equations with a sparse Cholesky decomposition. We consider convergence of the truncated Taylor approximation by studying band‐limited Matérn fields. We consider the convergence of the discrete approximations to the continuous limits. Finally, we study numerically the accuracy of different approximation methods with an interpolation problem.     

Multivariate Two-Sided Tests for Normal Mean Vectors Based on Approximations of Likelihood Ratio Test

Assuming that all components of a normal mean vector are simultaneously non negative or non positive, we consider a multivariate two-sided test for testing whether the normal mean vector is equal to zero or not. Since the likelihood ratio test is accompanied with theoretical and computational complications, we discuss two kinds of approximations of the likelihood ratio test. One is based on a conservative critical value determined by a certain inequality. The other is constructed by the approximation of the likelihood ratio test proposed by Tang et al. (1989). We compare the likelihood ratio test and two kinds of approximations through numerical examples regarding critical values and the power of the test.     

Density approximations and VaR computation for compound Poisson-lognormal distributions

Parametric approximations of the compound Poisson-lognormal distribution are developed and used to compute Value-at-Risk (VaR). As guidelines for finding an approximation, the skewness–kurtosis space and the tail behavior are considered. The Generalized Beta distribution of the second kind (GB2) and a mixture of lognormals are found to provide a good fit. In certain cases, the GB2 can be estimated by moment-matching, thus providing a simulation-free procedure for VaR computation. For confidence levels larger than 99%, extreme value theory approaches are developed. According to extensive Monte Carlo evidence, the proposed approximations are more efficient than crude Monte Carlo.     

LAN, LAM and convergence of Iterative ML estimates; optimal design in Gaussian one-way mixed model

For a one-way mixed Gaussian ANOVA model we prove local asymptotic normality and local asymptotic minimaxity of maximum likelihood estimates (MLE) and of its certain iterative approximations. A geometric rate of convergence in probability is proved for these iterative estimates to MLE. Asymptotically optimal designs for large samples are studied.     

A note on spearman's footrule

We obtain exact tables of the null distribution of Spearman's footrule for sample sizes n = 4(1)40 by using a certain Markov chain property, and we investigate the adequacy of approximations to the distribution.     

Simulation of Infinitely Divisible Random Fields

Two methods to approximate infinitely divisible random fields are presented. The methods are based on approximating the kernel function in the spectral representation of such fields, leading to numerical integration of the respective integrals. Error bounds for the approximation error are derived and the approximations are used to simulate certain classes of infinitely divisible random fields.     

More Efficient Approximation of Multiple Integrals using Steady State Ranked Simulated Sampling

This article extends the concept of using the steady state ranked simulated sampling approach (SRSIS) by Al-Saleh and Samawi (2000) for improving Monte Carlo methods for single integration problem to multiple integration problems. We demonstrate that this approach provides unbiased estimators and substantially improves the performance of some Monte Carlo methods for bivariate integral approximations, which can be extended to multiple integrals’ approximations. This results in a significant reduction in costs and time required to attain a certain level of accuracy. In order to compare the performance of our method with the Samawi and Al-Saleh (2007) method, we use the same two illustrations for the bivariate case.     

On the relative accuracy of certain bootstrap procedures

We show the second-order relative accuracy, on bounded sets, of the Studentized bootstrap, exponentially tilted bootstrap and nonparametric likelihood tilted bootstrap, for means and smooth functions of means. We also consider the relative errors for larger deviations. Our method exploits certain connections between Edgeworth and saddlepoint approximations to simplify the computations.     

Approximations for weighted Kolmogorov–Smirnov distributions via boundary crossing probabilities

A statistical application to Gene Set Enrichment Analysis implies calculating the distribution of the maximum of a certain Gaussian process, which is a modification of the standard Brownian bridge. Using the transformation into a boundary crossing problem for the Brownian motion and a piecewise linear boundary, it is proved that the desired distribution can be approximated by an n-dimensional Gaussian integral. Fast approximations are defined and validated by Monte Carlo simulation. The performance of the method for the genomics application is discussed.     

Saddlepoint approximations for nonlinear statistics

It is well known that saddlepoint expansions lead to accurate approximations to the cumulative distributions and densities of a sample mean and other simple linear statistics. The use of such expansions is explored in a broader situation. The saddlepoint formula for the tail probability of a certain type of nonlinear statistic is derived. The relative error of O(n^–1), as in the linear case, is retained. A simple example is considered, to illustrate the great accuracy of the approximation.     

Approximations to the distribution of weighted combination of independent probabilities

Two simple approximations are proposed for the distribution of the weighted combina-tion of n independent probabilities. The approximations are compared with other avail-able approximations. It is shown that one of the proposed approximations is better than the other approximations.     

Consistency of Generalized Maximum Spacing Estimates

General methods for the estimation of distributions can be derived from approximations of certain information measures. For example, both the maximum likelihood (ML) method and the maximum spacing (MSP) method can be obtained from approximations of the Kullback–Leibler information. The ideas behind the MSP method, whereby an estimation method for continuous univariate distributions is obtained from an approximation based on spacings of an information measure, were used by Ranneby & Ekstrom (1997) (using simple spacings) and Ekstrom (1997b) (using high order spacings) to obtain a class of methods, called generalized maximum spacing (GMSP) methods. In the present paper, GMSP methods will be shown to give consistent estimates under general conditions, comparable to those of Bahadur (1971) for the ML method, and those of Shao & Hahn (1999) for the MSP method. In particular, it will be proved that GMSP methods give consistent estimates in any family of distributions with unimodal densities, without any further conditions on the distributions.     

On the enumeration of rectangular (0, 1)-matrices

The construction and enumeration of (0, 1)-matrices with given line-sums is described for the rectangular cases often encountered in applications. Improved approximations are provided for the number of such matrices. Some new enumeration results for semi-regular bipartite graphs are included, and the related category of the quasi-semiregular bipartite graphs is recognized. The range of certain elements of products of a (0, I)-matrix is considered as a function of the line-sums. This, in turn, is related to the range in the numbers of interchanges available. Improvements in statistical practice that come from these constructions and enumerations are described.     

A note on the asymptotics of lattice paths with general boundaries

Lattice paths which are restricted by one or two nonlinear boundaries are of increasing importance in many applications of lattice path combinatorics, for instance in sequential statistics. It is typical for such applications that the number of paths which avoid certain boundaries has to be determined when the number of steps is large. Counting results, if available, and recursion are not very well suited in these cases. In this paper we present some asymptotic approximations which are easy to calculate and provide an accuracy which is sufficient for many practical purposes.