Bayesian testing of an exponential point null hypothesis

The problem of testing a point null hypothesis involving an exponential mean is The problem of testing a point null hypothesis involving an exponential mean is usual interpretation of P-values as evidence against precise hypotheses is faulty. As in Berger and Delampady (1986) and Berger and Sellke (1987), lower bounds on Bayesian measures of evidence over wide classes of priors are found emphasizing the conflict between posterior probabilities and P-values. A hierarchical Bayes approach is also considered as an alternative to computing lower bounds and “automatic” Bayesian significance tests which further illustrates the point that P-values are highly misleading measures of evidence for tests of point null hypotheses.     

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.     

An approach to evaluating sensitivity in Bayesian regression analyses

This paper presents a method for assessing the sensitivity of predictions in Bayesian regression analyses. In parametric Bayesian analyses there is a family s₀ of regression functions, parametrized by a finite-dimensional vector B. The family s₀ is a subset of R, the set of all possible regression functions. A prior π₀ on B induces a prior on R. This paper assesses sensitivity by computing bounds on the predictive probability of a fixed set K over a class of priors, Γ, induced by a class of families of regression functions, Γ_s, and a class of priors, Γ_π. This paper is divided into three parts which (1) define Γ, (2) describe an algorithm for finding accurate bounds on predictive probabilities over Γ and (3) illustrate the method with two examples. It is found that sensitivity to the family of regression functions can be much more important than sensitivity to π₀.     

Consistency of the kernel density estimator: a survey

Various consistency proofs for the kernel density estimator have been developed over the last few decades. Important milestones are the pointwise consistency and almost sure uniform convergence with a fixed bandwidth on the one hand and the rate of convergence with a fixed or even a variable bandwidth on the other hand. While considering global properties of the empirical distribution functions is sufficient for strong consistency, proofs of exact convergence rates use deeper information about the underlying empirical processes. A unifying character, however, is that earlier and more recent proofs use bounds on the probability that a sum of random variables deviates from its mean.     

Sharp bounds on l-estimates and their expectations for dependent samples

Sharp bounds on the expectation of each L-estimate are determined for samples of identically distributed random variables with a given expectation and a central absolute moment of order p > 1, The distributions achieving these bounds are constructed. Analogous results are derived in the cases of symmetric, noraiegative and bounded random variables, Accurate bounds for L-estimates of arbitrary finite samples are concluded.     

Zero-Bias Locally Adaptive Density Estimators

Strategies for improving fixed non-negative kernel estimators have focused on reducing the bias, either by employing higher-order kernels or by adjusting the bandwidth locally. Intuitively, bandwidths in the tails should be relatively larger in order to reduce wiggles since there is less data available in the tails. We show that in regions where the density function is convex, it is theoretically possible to find local bandwidths such that the pointwise bias is exactly zero. The corresponding pointwise mean squared error converges at the parametric rate of O ( n ⁻¹ ) rather than the slower O ( n ^−4/5). These so-called zero-bias bandwidths are constant and are usually orders of magnitude larger than the optimal locally adaptive bandwidths predicted by asymptotic mean squared error analysis. We describe data-based algorithms for estimating zero-bias bandwidths over intervals where the density is convex. We find that our particular density estimator attains the usual O ( n ^−4/5) rate. However, we demonstrate that the algorithms can provide significant improvement in mean squared error, often clearly visually superior curves, and a new operating point in the usual bias-variance tradeoff.     

Sharp Negative Bounds On Differences of Small Quantiles and Means of Distributions from Nonparametric Families

For the distributions with decreasing density and decreasing failure rate, and decreasing density and decreasing failure rate on the average, the quantiles are less than the mean if their orders do not exceed fixed levels. We determine the sharp negative upper bounds on the differences of the small quantiles and means for the distributions from the respective families in terms of scale units generated by pth absolute central moments. We show that the bounds amount trivially to zero when p > 1, and are strictly negative for p = 1.     

Variable sampling intervals for multiparameter shewhart charts

This paper considers the problem of using control charts to simultaneously monitor more than one parameter with emphasis on simultaneously monitoring the mean and variance. Fixed sampling interval control charts are modified to use variable sampling intervals depending on what is being observed from the data. Two basic strategies are investigated. One strategy uses separate control charts for each parameter, A second strategy uses a proposed single combined statistic which is sensitive to shifts in both the mean and variance. Each procedure is compared to corresponding fixed interval procedures. It is seen that for both strategies the variable sampling interval approach is substantially more efficient than fixed interval procedures.     

Combining Multivariate Bioassays: Accurate Inference Using Small Sample Asymptotics

For several independent multivariate bioassays performed at different laboratories or locations, the problem of testing the homogeneity of the relative potencies is addressed, assuming the usual slope‐ratio or parallel line assay model. When the homogeneity hypothesis holds, interval estimation of the common relative potency is also addressed. These problems have been investigated in the literature using likelihood‐based methods, under the assumption of a common covariance matrix across the different studies. This assumption is relaxed in this investigation. Numerical results show that the usual likelihood‐based procedures are inaccurate for both of the above problems, in terms of providing inflated type I error probabilities for the homogeneity test, and providing coverage probabilities below the nominal level for the interval estimation of the common relative potency, unless the sample sizes are large, as expected. Correction based on small sample asymptotics is investigated in this article, and this provides significantly more accurate results in the small sample scenario. The results are also illustrated with examples.     

Asymptotic normality of the lengths of a class of nonparametric confidence intervals for a regression parameter

In the linear regression model, the asymptotic distributions of certain functions of confidence bounds of a class of confidence intervals for the regression parameter arc investigated. The class of confidence intervals we consider in this paper are based on the usual linear rank statistics (signed as well as unsigned). Under suitable assumptions, if the confidence intervals are based on the signed linear rank statistics, it is established that the lengths, properly normalized, of the confidence intervals converge in law to the standard normal distributions; if the confidence intervals arc based on the unsigned linear rank statistics, it is then proved that a linear function of the confidence bounds converges in law to a normal distribution.     

Comparison of Bayesian sample size criteria: ACC, ALC, and WOC

A challenge for implementing performance-based Bayesian sample size determination is selecting which of several methods to use. We compare three Bayesian sample size criteria: the average coverage criterion (ACC) which controls the coverage rate of fixed length credible intervals over the predictive distribution of the data, the average length criterion (ALC) which controls the length of credible intervals with a fixed coverage rate, and the worst outcome criterion (WOC) which ensures the desired coverage rate and interval length over all (or a subset of) possible datasets. For most models, the WOC produces the largest sample size among the three criteria, and sample sizes obtained by the ACC and the ALC are not the same. For Bayesian sample size determination for normal means and differences between normal means, we investigate, for the first time, the direction and magnitude of differences between the ACC and ALC sample sizes. For fixed hyperparameter values, we show that the difference of the ACC and ALC sample size depends on the nominal coverage, and not on the nominal interval length. There exists a threshold value of the nominal coverage level such that below the threshold the ALC sample size is larger than the ACC sample size, and above the threshold the ACC sample size is larger. Furthermore, the ACC sample size is more sensitive to changes in the nominal coverage. We also show that for fixed hyperparameter values, there exists an asymptotic constant ratio between the WOC sample size and the ALC (ACC) sample size. Simulation studies are conducted to show that similar relationships among the ACC, ALC, and WOC may hold for estimating binomial proportions. We provide a heuristic argument that the results can be generalized to a larger class of models.     

Optimal bounds on l-statistics based on samples drawn with replacement from finite populations

We describe a method of establishing optimal bounds on the expectations of arbitrary linear combinations of order statistics based on iid samples drawn with replacement from finite populations of a fixed size. The bounds are expressed in terms of the population size, mean, central absolute moments, and coefficients of the combination. The bounds are precisely determined for the trimmed means and their differences, and single order statistics and their differences in particular. We also show that with increase in population size, our bounds approach the respective universal ones for arbitrary iid samples.     

Effects of correlated disturbances of some regression problems

In a linear regression model the disturbances are assumed to be independently distributed. If the correlation among the disturbances exists, then the usual F statistics have not the F distribution and the distributions depend on the regressor variables. This paper gives bounds on the F statistics, whose distributions do not depends on the regressor variables. The bounds are applied to a test on a general linear hypothesis of the regression coefficients and to evaluate the confidence level of a prediction set.     

Confidence Bounds for Multiplicative Comparisons

Statements that are inherently multiplicative have historically been justified using ratios of random variables. Although recent work on ratios has extended the classical theory to produce confidence bounds conditioned on a positive denominator, this current article offers a novel perspective that eliminates the need for such a condition. Although seemingly trivial, this new perspective leads to improved lower confidence bounds to support multiplicative statements. This perspective is also more satisfying as it allows comparisons that are inherently multiplicative in nature to be properly analyzed as such.     

The rate of convergence to stationarity for M/G/1 models with admission controls via coupling

We study the workload processes of two M/G/1 queueing systems with restricted capacity: in Model 1 any service requirement that would exceed a certain capacity threshold is truncated; in Model 2 new arrivals do not enter the system if they have to wait more than a fixed threshold time in line. For Model 1 we obtain several results concerning the rate of convergence to equilibrium. In particular, we derive uniform bounds for geometric ergodicity with respect to certain subclasses. For Model 2 geometric ergodicity follows from the finiteness of the moment-generating function of the service time distribution. We derive bounds for the convergence rates in special cases. The proofs use the coupling method.     

On confidence bounds for one-parameter exponential families

There exist various methods for providing confidence intervals for unknown parameters of interest on the basis of a random sample. Generally, the bounds are derived from a system of non-linear equations. In this article, we present a general solution to obtain an unbiased confidence interval with confidence coefficient 1 ? α in one-parameter exponential families. Also we discuss two Bayesian credible intervals, the highest posterior density (HPD) and relative surprise (RS) credible intervals. Standard criteria like the coverage length and coverage probability are used to assess the performance of the HPD and RS credible intervals. Simulation studies and real data applications are presented for illustrative purposes.     

Structural probability bounds for the strong Pareto law

By using the structural density function (Fraser 1979, Ch. 7) of the parameters of a Pareto distribution, the structural distribution function of the strong Pareto law is derived. Its fractiles have been evaluated numerically for special cases, and the results are displayed through graphs from which structural one-sided probability bounds may be found. It is shown that these graphs may also be used to find structural tolerance bounds for the Pareto distribution.     

Economic and economic–statistical designs of VSI Bayesian control chart using Monte Carlo method and ABC algorithm

The usual practice in using a Bayesian control chart to monitor a process is done by taking samples from the process with fixed sampling intervals. Recent studies on traditional control charts have shown that variable sampling interval (VSI) scheme compared to classical scheme (fixed ratio sampling, FRS) helps practitioners to detect process shifts more quickly. In this paper, the effectiveness of VSI scheme on performance of Bayesian control chart has been studied, based on economic (ED) and economic–statistical designs (ESD). Monte Carlo method and artificial bee colony algorithm have been utilized to obtain optimal design parameters of Bayesian control chart (sample size, sampling intervals, warning limit and control limit) since the statistic of this approach does not have any specified distribution. Finally, VSI Bayesian control chart has been compared to FRS Bayesian and VSI X-bar approaches based on ED and ESD, separately. According to the results, it has been found that the performance of VSI Bayesian scheme is better than FRS Bayesian and VSI X-bar approaches.     

Approximate Bayesianity of Frequentist Confidence Intervals for a Binomial Proportion

The well-known Wilson and Agresti–Coull confidence intervals for a binomial proportion p are centered around a Bayesian estimator. Using this as a starting point, similarities between frequentist confidence intervals for proportions and Bayesian credible intervals based on low-informative priors are studied using asymptotic expansions. A Bayesian motivation for a large class of frequentist confidence intervals is provided. It is shown that the likelihood ratio interval for p approximates a Bayesian credible interval based on Kerman’s neutral noninformative conjugate prior up to O(n^{? 1}) in the confidence bounds. For the significance level α ? 0.317, the Bayesian interval based on the Jeffreys’ prior is then shown to be a compromise between the likelihood ratio and Wilson intervals. Supplementary materials for this article are available online.     

Least significant spacing for ‘one versus the rest’ normal populations

Consider sample means from k(≥2) normal populations where the variances and sample sizes are equal. The problem is to find the ‘least significant difference’ or ‘spacing’ (LSS) between the two largest means, so that if an observed spacing is larger we have confidence 1 - α that the population with largest sample mean also has the largest population mean.

When the variance is known it is shown that the maximum LSS occurs when k = 2, provided a < .2723. In other words, for any value of k we may use the usual (one-tailed) least significant difference to demonstrate that one population has a population mean greater than (or equal to) the rest.

When the variance is estimated bounds are obtained for the confidence which indicate that this last result is approximately correct.