Exact likelihood inference for two populations from two-parameter exponential distributions under joint Type-II censoring

In this article, we develop exact inference for two populations that have a two-parameter exponential distribution with the same location parameter and different scale parameters when Type-II censoring is implemented on the two samples in a combined manner. We obtain the conditional maximum likelihood estimators (MLEs) of the three parameters. We then derive the exact distributions of these MLEs along with their moment generating functions. Based on general entropy loss function, Bayesian study about the parameters is presented. Finally, some simulation results and an illustrative example are presented to illustrate the methods of inference developed here.     

Kuiper's P-value as a measuring tool and decision procedure for the goodness-of-fit test

The need for computing P-values for the Kuiper statistic has been emphasised by Batschelet (1981). Some exact P-values, with useful interpretations for making inference about a probability model for circular data, are provided. Computation of the exact values are based on Durbin (1973) boundary crossing probabilities. A numerical example is used to demonstrate the usefulness of the results.     

A Bayesian adaptive design for two-stage clinical trials with survival data

A randomized two-stage adaptive Bayesian design is proposed and studied for allocation and comparison in a phase III clinical trial with survival time as treatment response. Several exact and limiting properties of the design and the follow-up inference are studied, both numerically and theoretically, and are compared with a single-stage randomized procedure. The applicability of the proposed methodology is illustrated by using some real data.     

THE GROUPING PROBLEM IN DISTRIBUTION-FREE PLANAR REGRESSION

In the planar regression model having two slope parameters and identically distributed errors, exact distribution-free inference about one parameter may be carried out by grouping the observations, eliminating the nuisance parameter and reducing the model to simple linear regression, allowing exact distribution-free methods for slope to be employed. This model reduction involves a loss of efficiency: the choice of an optimal grouping to minimize efficiency loss is discussed.     

Robustness of Inference for One-sample Problem with Correlated Observations

The inference about the population mean based on the standard t-test involves the assumption of normal population as well as independence of the observations. In this paper we examine the robustness of the inference in the presence of correlations among the observations. We consider the simplest correlation structure AR(1) and its impact on the t-test. A modification of the t-test suitable for this structure is suggested, and its effect on the inference is investigated using Monte Carlo simulation.     

An efficient numerical algorithm for exact inference in meta analysis

The performance of commonly used asymptotic inference procedures for the random-effects model used in meta analysis relies on the number of studies. When the number of studies is moderate or small, the exact inference procedure is more reliable than the asymptotic counterparts. However, the related numerical computation may be demanding and an obstacle of routine use of the exact method. In this paper, we proposed a novel numerical algorithm for constructing the exact 95% confidence interval of the location parameter in the random-effects model. The algorithm is much faster than the naive method and may greatly facilitate the use of the more appropriate exact inference procedure in meta analysis. Numerical studies and real data examples are used to illustrate the advantage of the proposed method.     

Bayesian inference for epidemics with two levels of mixing

Abstract.  Methodology for Bayesian inference is considered for a stochastic epidemic model which permits mixing on both local and global scales. Interest focuses on estimation of the within- and between-group transmission rates given data on the final outcome. The model is sufficiently complex that the likelihood of the data is numerically intractable. To overcome this difficulty, an appropriate latent variable is introduced, about which asymptotic information is known as the population size tends to infinity. This yields a method for approximate inference for the true model. The methods are applied to real data, tested with simulated data, and also applied to a simple epidemic model for which exact results are available for comparison.     

Exact Inference in the Proportional Hazard Model: Possibilities and Limitations

It is suggested that inference under the proportional hazard model can be carried out by programs for exact inference under the logistic regression model. Advantages of such inference is that software is available and that multivariate models can be addressed. The method has been evaluated by means of coverage and power calculations in certain situations. In all situations coverage was above the nominal level, but on the other hand rather conservative. A different type of exact inference is developed under Type II censoring. Inference was then less conservative, however there are limitations with respect to censoring mechanism, multivariate generalizations and software is not available. This method also requires extensive computational power. Performance of large sample Wald, score and likelihood inference was also considered. Large sample methods works remarkably well with small data sets, but inference by score statistics seems to be the best choice. There seems to be some problems with likelihood ratio inference that may originate from how this method works with infinite estimates of the regression parameter. Inference by Wald statistics can be quite conservative with very small data sets.     

Exact Inference for a Population Proportion Based on a Ranked Set Sample

ABSTRACT

This article develops and investigates a confidence interval and hypothesis testing procedure for a population proportion based on a ranked set sample (RSS). The inference is exact, in the sense that it is based on the exact distribution of the total number of successes observed in the RSS. Furthermore, this distribution can be readily computed with the well-known and freely available R statistical software package. A data example that illustrates the methodology is presented. In addition, the properties of the inference procedures are compared with their simple random sample (SRS) counterparts. In regards to expected lengths of confidence intervals and the power of tests, the RSS inference procedures are superior to the SRS methods.     

GROUPING PROBLEMS IN DISTRIBUTION-FREE REGRESSION1

In regression models having symmetric errors, exact distribution-free inference about individual parameters may be carried out by grouping observations, eliminating unwanted parameters within groups, and applying distribution free techniques for the symmetric location parameter problem. Models whose errors have identical but not symmetric distributions may obtain symmetry by taking differences between pairs of observations. Both grouping and differencing involve potential efficiency loss. The choice of an optimal scheme to minimize efficiency loss is expressible as a multi–assignment type of problem, whose solutions, exact and approximate, are discussed.     

Exact inference for competing risks model with generalized type-I hybrid censored exponential data

Recently, exact inference under hybrid censoring scheme has attracted extensive attention in the field of reliability analysis. However, most of the authors neglect the possibility of competing risks model. This paper mainly discusses the exact likelihood inference for the analysis of generalized type-I hybrid censoring data with exponential competing failure model. Based on the maximum likelihood estimates for unknown parameters, we establish the exact conditional distribution of parameters by conditional moment generating function, and then obtain moment properties as well as exact confidence intervals (CIs) for parameters. Furthermore, approximate CIs are constructed by asymptotic distribution and bootstrap method as well. We also compare their performances with exact method through the use of Monte Carlo simulations. And finally, a real data set is analysed to illustrate the validity of all the methods developed here.     

Accelerating inference for diffusions observed with measurement error and large sample sizes using approximate Bayesian computation

In recent years, dynamical modelling has been provided with a range of breakthrough methods to perform exact Bayesian inference. However, it is often computationally unfeasible to apply exact statistical methodologies in the context of large data sets and complex models. This paper considers a nonlinear stochastic differential equation model observed with correlated measurement errors and an application to protein folding modelling. An approximate Bayesian computation (ABC)-MCMC algorithm is suggested to allow inference for model parameters within reasonable time constraints. The ABC algorithm uses simulations of ‘subsamples’ from the assumed data-generating model as well as a so-called ‘early-rejection’ strategy to speed up computations in the ABC-MCMC sampler. Using a considerate amount of subsamples does not seem to degrade the quality of the inferential results for the considered applications. A simulation study is conducted to compare our strategy with exact Bayesian inference, the latter resulting two orders of magnitude slower than ABC-MCMC for the considered set-up. Finally, the ABC algorithm is applied to a large size protein data. The suggested methodology is fairly general and not limited to the exemplified model and data.     

Markov chain Monte Carlo exact inference for binomial and multinomial logistic regression models

We develop Metropolis-Hastings algorithms for exact conditional inference, including goodness-of-fit tests, confidence intervals and residual analysis, for binomial and multinomial logistic regression models. We present examples where the exact results, obtained by enumeration, are available for comparison. We also present examples where Monte Carlo methods provide the only feasible approach for exact inference.     

Exact inference and prediction for k-sample exponential case under type-ii censoring

The exact inference and prediction intervals for the K-sample exponential scale parameter under doubly Type-II censored samples are derived using an algorithm of Huffer and Lin [Huffer, F.W. and Lin, C.T., 2001, Computing the joint distribution of general linear combinations of spacings or exponen-tial variates. Statistica Sinica, 11, 1141–1157.]. This approach provides a simple way to determine the exact percentage points of the pivotal quantity based on the best linear unbiased estimator in order to develop exact inference for the scale parameter as well as to construct exact prediction intervals for failure times unobserved in the ith sample. Similarly, exact prediction intervals for failure times of units from a future sample can also be easily obtained.     

Exact or approximate inference in graphical models: why the choice is dictated by the treewidth,and how variable elimination can be exploited

Probabilistic graphical models offer a powerful framework to account for the dependence structure between variables, which is represented as a graph. However, the dependence between variables may render inference tasks intractable. In this paper, we review techniques exploiting the graph structure for exact inference, borrowed from optimisation and computer science. They are built on the principle of variable elimination whose complexity is dictated in an intricate way by the order in which variables are eliminated. The so‐called treewidth of the graph characterises this algorithmic complexity: low‐treewidth graphs can be processed efficiently. The first point that we illustrate is therefore the idea that for inference in graphical models, the number of variables is not the limiting factor, and it is worth checking the width of several tree decompositions of the graph before resorting to the approximate method. We show how algorithms providing an upper bound of the treewidth can be exploited to derive a ‘good' elimination order enabling to realise exact inference. The second point is that when the treewidth is too large, algorithms for approximate inference linked to the principle of variable elimination, such as loopy belief propagation and variational approaches, can lead to accurate results while being much less time consuming than Monte‐Carlo approaches. We illustrate the techniques reviewed in this article on benchmarks of inference problems in genetic linkage analysis and computer vision, as well as on hidden variables restoration in coupled Hidden Markov Models.     

Regularized Gaussian belief propagation

Belief propagation (BP) has been applied in a variety of inference problems as an approximation tool. BP does not necessarily converge in loopy graphs, and even if it does, is not guaranteed to provide exact inference. Even so, BP is useful in many applications due to its computational tractability. In this article, we investigate a regularized BP scheme by focusing on loopy Markov graphs (MGs) induced by a multivariate Gaussian distribution in canonical form. There is a rich literature surrounding BP on Gaussian MGs (labelled Gaussian belief propagation or GaBP), and this is known to experience the same problems as general BP on graphs. GaBP is known to provide the correct marginal means if it converges (this is not guaranteed), but it does not provide the exact marginal precisions. We show that our adjusted BP will always converge, with sufficient tuning, while maintaining the exact marginal means. As a further contribution we show, in an empirical study, that our GaBP variant can accelerate GaBP and compares well with other GaBP-type competitors in terms of convergence speed and accuracy of approximate marginal precisions. These improvements suggest that the principle of regularized BP should be investigated in other inference problems. The selection of the degree of regularization is addressed through the use of two heuristics. A by-product of GaBP is that it can be used to solve linear systems of equations; the same is true for our variant and we make an empirical comparison with the conjugate gradient method.     

An exact test for trend among binomial proportions based on a modified Baumgartner-Weiß-Schindler statistic

The Cochran-Armitage test is the most frequently used test for trend among binomial proportions. This test can be performed based on the asymptotic normality of its test statistic or based on an exact null distribution. As an alternative, a recently introduced modification of the Baumgartner-Weiß-Schindler statistic, a novel nonparametric statistic, can be used. Simulation results indicate that the exact test based on this modification is preferable to the Cochran-Armitage test. This exact test is less conservative and more powerful than the exact Cochran-Armitage test. The power comparison to the asymptotic Cochran-Armitage test does not show a clear winner, but the difference in power is usually small. The exact test based on the modification is recommended here because, in contrast to the asymptotic Cochran-Armitage test, it guarantees a type I error rate less than or equal to the significance level. Moreover, an exact test is often more appropriate than an asymptotic test because randomization rather than random sampling is the norm, for example in biomedical research. The methods are illustrated with an example data set.     

On the foundations of likelihood principle

According to the likelihood principle, if the designs produce proportional likelihood functions, one could make an identical inference about a parameter from the data irrespective of the design, which yields the data. If it comes to that, there are several counter-examples, and/or paradoxical consequences to likelihood principle. Besides, as we will see, contrary to a widely held opinion, such a principle is not a direct consequence of Bayes theorem. In particular, the piece of information about the design is one part of the evidence, and it is relevant for the prior. Later on, Jeffreys non-informative prior is used to show how different designs result in different priors. Another basic idea of the present paper is that (apart from other information) the equiprobability assumption is to be linked to the idea of the impartiality of design with respect to the parameter under consideration. The whole paper has remarkable implications on the foundations of statistics from the notion of sufficiency, the relevance of the stopping rule and of the randomization in survey sampling and in the experimental design, the difference between ignorable and non-ignorable designs, until a reconciliation of different approaches to the inductive reasoning in statistical inference.     

BAYESIAN ANALYSIS OF A FRACTIONAL COINTEGRATION MODEL

The concept of fractional cointegration, whereby deviations from an equilibrium relationship follow a fractionally integrated process, has attracted some attention of late. The extended concept allows cointegration to be associated with mean reversion in the error, rather than requiring the more stringent condition of stationarity. This paper presents a Bayesian method for conducting inference about fractional cointegration. The method is based on an approximation of the exact likelihood, with a Jeffreys prior being used to offset identification problems. Numerical results are produced via a combination of Markov chain Monte Carlo algorithms. The procedure is applied to several purchasing power parity relations, with substantial evidence found in favor of parity reversion.     

Application of Bayesian analyses to doubly randomized delayed start,matched control designs to demonstrate disease modification

Disease modification is a primary therapeutic aim when developing treatments for most chronic progressive diseases. The best treatments do not simply affect disease symptoms but fundamentally improve disease course by slowing, halting, or reversing disease progression. One of many challenges for establishing disease modification relates to the identification of adequate analytic tools to show differences in a disease course following intervention. Traditional approaches rely on the comparisons of slopes or noninferiority margins. However, it has proven difficult to conclusively demonstrate disease modification using such approaches. To address these challenges, we propose a novel adaptation of the delayed start study design that incorporates posterior probabilities identified by hierarchical Bayesian inference approaches to establish evidence for disease modification. Our models compare the size of treatment differences at the end of the delayed start period with those at the end of the early start period. Simulations that compare several models are provided. These include general linear models, repeated measures models, spline models, and model averaging. Our work supports the superiority of model averaging for accurately characterizing complex data that arise in real world applications. This novel approach has been applied to the design of an ongoing, doubly randomized, matched control study that aims to show disease modification in young persons with schizophrenia (the Disease Recovery Evaluation and Modification (DREaM) study). The application of this Bayesian methodology to the DREaM study highlights the value of this approach and demonstrates many practical challenges that must be addressed when implementing this methodology in a real world trial.