Universal domination of stein-type estimators

For the problem of estimating the location parameter of a p-variate spherically symmetric distribution (p>3), Hwang (1985) established the dominance of some positive-part James-Stein (1961) estimators over the usual estimator simultaneously under a very general class of loss function. Vie show that many of his results can be extended to a class of positive-part Baranchik-type estimators (1970).     

A Note on the Use of Confidence Limits following Rejection of a Null Hypothesis

The equivalence of some tests of hypothesis and confidence limits is well known. When, however, the confidence limits are computed only after rejection of a null hypothesis, the usual unconditional confidence limits are no longer valid. This refers to a strict two-stage inference procedure: first test the hypothesis of interest and if the test results in a rejection decision, then proceed with estimating the relevant parameter. Under such a situation, confidence limits should be computed conditionally on the specified outcome of the test under which estimation proceeds. Conditional confidence sets will be longer than unconditional confidence sets and may even contain values of the parameter previously rejected by the test of hypothesis. Conditional confidence limits for the mean of a normal population with known variance are used to illustrate these results. In many applications, these results indicate that conditional estimation is probably not good practice.     

JAMES-STEIN RULE ESTIMATORS IN LINEAR REGRESSION MODELS WITH MULTIVARIATE-t DISTRIBUTED ERROR

This paper considers estimation of β in the regression model y =Xβ+μ, where the error components in μ have the jointly multivariate Student-t distribution. A family of James-Stein type estimators (characterised by nonstochastic scalars) is presented. Sufficient conditions involving only X are given, under which these estimators are better (with respect to the risk under a general quadratic loss function) than the usual minimum variance unbiased estimator (MVUE) of β. Approximate expressions for the bias, the risk, the mean square error matrix and the variance-covariance matrix for the estimators in this family are obtained. A necessary and sufficient condition for the dominance of this family over MVUE is also given.     

Second-order inference for generalized least squares

Confidence regions for generalized least squares are commonly derived from a measure of departure calculated on the tangent plane at the MLE or on the tangent plane at the true value; the first gives approximate confidence regions, the second exact. For surfaces with curvature, indeed with varying curvature, the exact regions typically are not likelihood regions and can include parameter values of highest and of lowest likelihood. This paper develops an alternative approach to deriving exact confidence regions and uses both surface curvature and distance from the surface as supporting ingredients. For this, conditionality is invoked in two ways beyond that supported by the usual conditionality principle. For the case of normal error the ordinary chi-squared departure is replaced by a Von Mises-type angular (or cosine) departure which is assessed using curvature properties in the data direction and radial distance of the data from the regression surface. For the usual linear model (constant curvature equal to zero) the method coincides with the ordinary tests and confidence regions; for the case of constant nonzero curvature, the method generalizes to spheres and sphere-cylinders the Fisher (Statistical Methods and Scientific Inference, 1956) analysis of a rotationally symmetric normal on ?² with mean constrained to a circle. The effects of conditioning are indicated by a computer plot for obtaining 95% confidence.     

Limit expressions for the risk of james-stein estimators

Limit expressions (as the dimension p ← ∞ ) are derived for the relative risk of the James-Stein estimator and its positive-part version. The limit is simple to evaluate, and gives the amount of improvement in risk that is possible. The technique used is to bound the risk, both above and below. with bounds that converge to the same limit. For the James-Stein estimator these bounds are simple to calculate, and are quite accurate even for moderate dimensions.     

Confidence sets based on the positive part James–Stein estimator with the asymptotically constant coverage probability

The asymptotic expansions for the coverage probability of a confidence set centred at the James–Stein estimator presented in our previous publications show that this probability depends on the non-centrality parameter τ² (the sum of the squares of the means of normal distributions). In this paper we establish how these expansions can be used for a construction of confidence region with constant confidence level, which is asymptotically (the same formula for both case τ→0 and τ→∞) equal to some fixed value 1?α. We establish the shrinkage rate for the confidence region according to the growth of the dimension p and also the value of τ for which we observe quick decreasing of the coverage probability to the nominal level 1?α. When p→∞ this value of τ increases as O(p^1/4). The accuracy of the results obtained is shown by the Monte-Carlo statistical simulations.     

Conditional interval estimation of the mean following rejection of a two sided test

The effect of rejecting a two-sided preliminary test of significance for the mean of a normal distribution upon subsequent interval estimation of the mean is examined. For the case where the variance is known, conditional confidence intervals may be shorter than unconditional intervals, in contrast to the one-sided preliminary test case examined by Meeks and D’Agostino (1983, The American Statistician, 7, 134-136) . For the case where the variance is unknown and must be estimated by the sample variance, it is shown that customary intervals do not offer uniformly greater or lesser coverage than the nominal level.     

The two-sided power distribution for the treatment of the uncertainty in PERT

The aim of this paper is to include the Two-Sided Power (TSP) distribution in the PERT methodology making use of the advantages that this four-parameter distribution offers. In order to be completely determined, a distribution of this type needs, the same as the beta distribution, a new datum apart from the three usual values a (pessimistic), m (most likely) and b (optimistic). To solve this question, when using the beta distribution in the PERT context, we are looking for the maximum similarity with the normal and so it is required that the distribution has the same variance as the normal or its same kurtosis, giving rise to the constant variance and mesokurtic families, respectively. Nevertheless, while this approach can be only applied to the beta distribution for some values in the range of the standardized mode, in the case of the TSP distribution this methodology leads always to a solution. A detailed analysis comparing the beta and TSP distribution based on their PERT means and variances is presented indicating better results for the second. We are very grateful for the comments and suggestions of two anonymous referees.     

Improved Score Tests in Symmetric Linear Regression Models

The class of symmetric linear regression models has the normal linear regression model as a special case and includes several models that assume that the errors follow a symmetric distribution with longer-than-normal tails. An important member of this class is the t linear regression model, which is commonly used as an alternative to the usual normal regression model when the data contain extreme or outlying observations. In this article, we develop second-order asymptotic theory for score tests in this class of models. We obtain Bartlett-corrected score statistics for testing hypotheses on the regression and the dispersion parameters. The corrected statistics have chi-squared distributions with errors of order O(n ^?3/2), n being the sample size. The corrections represent an improvement over the corresponding original Rao's score statistics, which are chi-squared distributed up to errors of order O(n ^?1). Simulation results show that the corrected score tests perform much better than their uncorrected counterparts in samples of small or moderate size.     

A comparison of estimators for the proportion in the tail of a normal distribution

Three estimators of the proportion in a tail of the normal distribution are compared using the criteria of mean squared error and mean absolute error. The estimators that we compare are the maximum likelihood estimator, the minimum variance unbiased estimator, and an intuitive estimator that is frequently used in practice. The intuitive estimator is similar to the MLE but uses the usual unbiased estimator of σ² rather than the MLE of σ². We show that the intuitive estimator has low efficiency, and for this reason it is not recommended. For very smallp and for largep the MVUE has the highest efficiency. The MLE is best for moderate values ofp.     

Expansions of the coverage probabilities of prediction region based on a shrinkage estimator

An expansion formula for the coverage probability of prediction region based on a shrinkage estimator proposed by Joshi [Joshi, V. M. (1967). Inadmissibility of the usual confidence sets for the mean of a multivariate normal population. Ann. Math. Statist., 38, 1868–1875.] is obtained. Its error bound is evaluated in terms of a function of an unknown parameter. Applying this result, three types of asymptotic expansions are derived. These expansions show inadmissibility of the usual prediction region.     

On elliptical multilevel models

Multilevel models have been widely applied to analyze data sets which present some hierarchical structure. In this paper we propose a generalization of the normal multilevel models, named elliptical multilevel models. This proposal suggests the use of distributions in the elliptical class, thus involving all symmetric continuous distributions, including the normal distribution as a particular case. Elliptical distributions may have lighter or heavier tails than the normal ones. In the case of normal error models with the presence of outlying observations, heavy-tailed error models may be applied to accommodate such observations. In particular, we discuss some aspects of the elliptical multilevel models, such as maximum likelihood estimation and residual analysis to assess features related to the fitting and the model assumptions. Finally, two motivating examples analyzed under normal multilevel models are reanalyzed under Student-t and power exponential multilevel models. Comparisons with the normal multilevel model are performed by using residual analysis.     

Improved robust Bayes estimators of the error variance in linear models

We consider the problem of estimating the error variance in a general linear model when the error distribution is assumed to be spherically symmetric, but not necessary Gaussian. In particular we study the case of a scale mixture of Gaussians including the particularly important case of the multivariate-t distribution. Under Stein's loss, we construct a class of estimators that improve on the usual best unbiased (and best equivariant) estimator. Our class has the interesting double robustness property of being simultaneously generalized Bayes (for the same generalized prior) and minimax over the entire class of scale mixture of Gaussian distributions.     

CONDITIONAL INFERENCE ABOUT AN EQUICORRELATION COEFFICIENT

SenGupta (1987) proposed a locally most powerful test which is globally (one sided) unbiased, and an estimator of p, the equicorrelation coefficient of a standard symmetric multivariate normal (SSMN) distribution. Here we use the idea in Williams (1984) to illustrate the construction and use of ancillary statistics to make inference about p. The test and confidence intervals based on this construction are conditionally optimal.     

Hypothesis Testing and Confidence Regions for the Mean Sojourn Time of an M/M/1 Queueing System

This article discusses testing hypotheses and confidence regions with correct levels for the mean sojourn time of an M/M/1 queueing system. The uniformly most powerful unbiased tests for three usual hypothesis testing problems are obtained and the corresponding p values are provided. Based on the duality between hypothesis tests and confidence sets, the uniformly most accurate confidence bounds are derived. A confidence interval with correct level is proposed.     

Nonparametric confidence intervals for location in time series data

We extend the confidence interval construction procedure for location for symmetric iid data using the one-sample Wilcoxon signed rank statistic (T⁺) to stationary time series data. We propose a normal approximation procedure when explicit knowledge of the underlying dependence structure/distribution is unknown. By conducting extensive simulations from linear and nonlinear time series models, we show that the extended procedure is a strong contender for use in the construction of confidence intervals in time series analysis. Finally we demonstrate real application implementations in two case studies.     

Confidence Sets Based on Thresholding Estimators in High-Dimensional Gaussian Regression Models

We study confidence intervals based on hard-thresholding, soft-thresholding, and adaptive soft-thresholding in a linear regression model where the number of regressors k may depend on and diverge with sample size n. In addition to the case of known error variance, we define and study versions of the estimators when the error variance is unknown. In the known-variance case, we provide an exact analysis of the coverage properties of such intervals in finite samples. We show that these intervals are always larger than the standard interval based on the least-squares estimator. Asymptotically, the intervals based on the thresholding estimators are larger even by an order of magnitude when the estimators are tuned to perform consistent variable selection. For the unknown-variance case, we provide nontrivial lower bounds and a small numerical study for the coverage probabilities in finite samples. We also conduct an asymptotic analysis where the results from the known-variance case can be shown to carry over asymptotically if the number of degrees of freedom n ? k tends to infinity fast enough in relation to the thresholding parameter.     

Robustness of three power transformation estimation procedures

Maximum likelihood, goodness-of-fit, and symmetric percentile estimators of the power transformation parameterp, are considered. The comparative robustness of each estimation procedure is evaluated when the transformed data can be made symmetric, but may not necessarily be normal. Seven types of symmetric distributions are considered as well as four contaminated normal distributions over a range of six p values for samples of size 25, 50, and 100. The results indicate that the maximum likelihood estimator was slightly better than the goodness-of-fit estimator, but both were greatly superior to the percentile estimator. In general, the procedures were robust to distributional symmetric departures from normality, but increasing kurtosis caused appreciable increases in variation for estimated p values. The variability of p was found to decrease more than exponentially with decreases in the underlying normal distribution coefficient of variation. The standard likelihood ratio confidence interval procedure was found not to be generally useful.     

Expected Absolute Error of the Usual Estimator of the Binomial Parameter

For X with binomial (n, p) distribution the usual measure of the error of X/n as an estimator of p is its standard error S_n(p) = √{E(X/n – p)²} = √{p(1 – p)/n}. A somewhat more natural measure is the average absolute error D_n(p) = E‖X/n – p‖. This article considers use of D_n(p) instead of S_n(p) in a student's first introduction to statistical estimation. Exact and asymptotic values of D_n(p), and the appearance of its graph, are described in detail. The same is done for the Poisson distribution.     

A Note on the Comparison of the Stein Estimator and the James-Stein Estimator

The seminal work of Stein (1956 Stein, C. (1956). Inadmissibility of the usual estimator for the mean of a multivariate normal distribution. Proc. Third Berkeley Symp. Mathemat. Statist. Probab., University of California Press, 1:197–206. [Google Scholar]) showed that the maximum likelihood estimator (MLE) of the mean vector of a p-dimensional multivariate normal distribution is inadmissible under the squared error loss function when p ? 3 and proposed the Stein estimator that dominates the MLE. Later, James and Stein (1961 James, W., Stein, C. (1961). Estimation with quadratic loss. Proc. Fourth Berkeley Symp. Mathemat. Statist. Probab., University of California Press, 1:361–379. [Google Scholar]) proposed the James-Stein estimator for the same problem and received much more attention than the original Stein estimator. We re-examined the Stein estimator and conducted an analytic comparison with the James-Stein estimator. We found that the Stein estimator outperforms the James-Stein estimator under certain scenarios and derived the sufficient conditions.