Proper Posterior Distribution Research Articles

Objective Bayesian analysis for geostatistical Student-t processes

ABSTRACT The choice of the prior distribution is a key aspect of Bayesian analysis. The spatial Student-t regression model poses some challenges when eliciting priors. It is well known that the propriety of the posterior distribution over objective priors is not always guaranteed, whereas the use of proper prior distributions may dominate and bias the posterior analysis. In this paper, we show the conditions under which our proposed reference prior and the two introduced Jeffreys priors yield to a proper posterior distribution. Simulation studies and a real data application are used in order to evaluate the performance of the reference prior.

Bayesian Reference Analysis for the Generalized Normal Linear Regression Model

This article proposes the use of the Bayesian reference analysis to estimate the parameters of the generalized normal linear regression model. It is shown that the reference prior led to a proper posterior distribution, while the Jeffreys prior returned an improper one. The inferential purposes were obtained via Markov Chain Monte Carlo (MCMC). Furthermore, diagnostic techniques based on the Kullback–Leibler divergence were used. The proposed method was illustrated using artificial data and real data on the height and diameter of Eucalyptus clones from Brazil.

Objective Bayesian Analysis for Gaussian Hierarchical Models with Intrinsic Conditional Autoregressive Priors

Bayesian hierarchical models are commonly used for modeling spatially correlated areal data. However, choosing appropriate prior distributions for the parameters in these models is necessary and sometimes challenging. In particular, an intrinsic conditional autoregressive (CAR) hierarchical component is often used to account for spatial association. Vague proper prior distributions have frequently been used for this type of model, but this requires the careful selection of suitable hyperparameters. In this paper, we derive several objective priors for the Gaussian hierarchical model with an intrinsic CAR component and discuss their properties. We show that the independence Jeffreys and Jeffreys-rule priors result in improper posterior distributions, while the reference prior results in a proper posterior distribution. We present results from a simulation study that compares frequentist properties of Bayesian procedures that use several competing priors, including the derived reference prior. We demonstrate that using the reference prior results in favorable coverage, interval length, and mean squared error. Finally, we illustrate our methodology with an application to 2012 housing foreclosure rates in the 88 counties of Ohio.

A Bayesian approach to Mendelian randomization with multiple pleiotropic variants.

SummaryWe propose a Bayesian approach to Mendelian randomization (MR), where instruments are allowed to exert pleiotropic (i.e. not mediated by the exposure) effects on the outcome. By having these effects represented in the model by unknown parameters, and by imposing a shrinkage prior distribution that assumes an unspecified subset of the effects to be zero, we obtain a proper posterior distribution for the causal effect of interest. This posterior can be sampled via Markov chain Monte Carlo methods of inference to obtain point and interval estimates. The model priors require a minimal input from the user. We explore the performance of our method by means of a simulation experiment. Our results show that the method is reasonably robust to the presence of directional pleiotropy and moderate correlation between the instruments. One section of the article elaborates the model to deal with two exposures, and illustrates the possibility of using MR to estimate direct and indirect effects in this situation. A main objective of the article is to create a basis for developments in MR that exploit the potential offered by a Bayesian approach to the problem, in relation with the possibility of incorporating external information in the prior, handling multiple sources of uncertainty, and flexibly elaborating the basic model.

Bayesian methods for type A evaluation of standard uncertainty

One of the aims of the revision of the guide to the expression of uncertainty in measurement (GUM), is to make it self-consistent as the current edition would combine frequentist and Bayesian statistical methods. The methods for type A evaluation of standard uncertainty in the GUM are considered to be ‘frequentist’ and hence would have to be replaced by their Bayesian counterparts. The aim of this paper is to propose Bayesian methods for four mainstream type A evaluations of standard uncertainty, based on the same assumptions as those underlying the current methods for type A evaluation given in the GUM. Specific attention is given to the computational aspects of the Bayesian methods. An approach involving weakly informative prior distributions is proposed to ensure a proper posterior distribution, still allowing the data to dominate it. The applications for which Bayesian methods are proposed include the calculation of a mean of a series of observations, analysis of variance, ordinary least squares and errors-in-variables regression. Examples H.3 (ordinary least squares) and H.5 (analysis of variance) from the GUM are reworked accordingly.The second aim of this paper is to propose Bayesian methods for type A evaluation that take the information typically available in metrology into account in the form of weakly informative prior distributions. These methods can applied in a wide variety of situations. By comparing the results from the proposed Bayesian methods involving weakly informative priors with their classical counterparts, it is seen that in most instances these methods give very similar estimates and standard uncertainties as classical methods do under the same conditions and assumptions. Estimates are often identical up to the last meaningful digit, and the standard uncertainties are usually slightly yet persistently larger in comparison to its classical counterpart. In many cases these differences turn out to have limited or little practical meaning, save in the case of a small series of observations, where the differences in the computed standard uncertainty of the mean is larger.

Bayesian analysis of the generalized gamma distribution using non-informative priors

ABSTRACTThe Generalized gamma (GG) distribution plays an important role in statistical analysis. For this distribution, we derive non-informative priors using formal rules, such as Jeffreys prior, maximal data information prior and reference priors. We have shown that these most popular formal rules with natural ordering of parameters, lead to priors with improper posteriors. This problem is overcome by considering a prior averaging approach discussed in Berger et al. [Overall objective priors. Bayesian Analysis. 2015;10(1):189–221]. The obtained hybrid Jeffreys-reference prior is invariant under one-to-one transformations and yields a proper posterior distribution. We obtained good frequentist properties of the proposed prior using a detailed simulation study. Finally, an analysis of the maximum annual discharge of the river Rhine at Lobith is presented.

Incorporating unobserved heterogeneity in Weibull survival models: A Bayesian approach

Outlying observations and other forms of unobserved heterogeneity can distort inference for survival datasets. The family of Rate Mixtures of Weibull distributions includes subject-level frailty terms as a solution to this issue. With a parametric mixing distribution assigned to the frailties, this family generates flexible hazard functions. Covariates are introduced via an Accelerated Failure Time specification for which the interpretation of the regression coefficients does not depend on the choice of mixing distribution. A weakly informative prior is proposed by combining the structure of the Jeffreys prior with a proper prior on some model parameters. This improper prior is shown to lead to a proper posterior distribution under easily satisfied conditions. By eliciting the proper component of the prior through the coefficient of variation of the survival times, prior information is matched for different mixing distributions. Posterior inference on subject-level frailty terms is exploited as a tool for outlier detection. Finally, the proposed methodology is illustrated using two real datasets, one concerning bone marrow transplants and another on cerebral palsy.

BAYESIAN MULTISCALE FINITE ELEMENT METHODS. MODELING MISSING SUBGRID INFORMATION PROBABILISTICALLY

In this paper, we develop a Bayesian multiscale approach based on a multiscale finite element method. Because of scale disparity in many multiscale applications, computational models can not resolve all scales. Various subgrid models are proposed to represent un-resolved scales. Here, we consider a probabilistic approach for modeling un-resolved scales using the Multiscale Finite Element Method (cf., [1, 2]). By representing dominant modes using the Generalized Multiscale Finite Element, we propose a Bayesian framework, which provides multiple inexpensive (computable) solutions for a deterministic problem. These approximate probabilistic solutions may not be very close to the exact solutions and, thus, many realizations are needed. In this way, we obtain a rigorous probabilistic description of approximate solutions. In the paper, we consider parabolic and wave equations in heterogeneous media. In each time interval, the domain is divided into subregions. Using residual information, we design appropriate prior and posterior distributions. The likelihood consists of the residual minimization. To sample from the resulting posterior distribution, we consider several sampling strategies. The sampling involves identifying important regions and important degrees of freedom beyond permanent basis functions, which are used in residual computation. Numerical results are presented. We consider two sampling algorithms. The first algorithm uses sequential sampling and is inexpensive. In the second algorithm, we perform full sampling using the Gibbs sampling algorithm, which is more accurate compared to the sequential sampling. The main novel ingredients of our approach consist of: defining appropriate permanent basis functions and the corresponding residual; setting up a proper posterior distribution; and sampling the posteriors.

Posterior property of Student-[formula omitted] linear regression model using objective priors

We derive objective priors for linear model with independent Student-t errors with unknown degrees of freedom v. It is shown that most of them preclude the existence of a proper posterior distribution unless we are willing to truncate the possible values of va priori.

Posterior propriety in Bayesian extreme value analyses using reference priors

The Generalized Pareto (GP) and Generalized extreme value (GEV) distributions play an important role in extreme value analyses, as models for threshold excesses and block maxima respectively. For each of these distributions we consider Bayesian inference using "reference" prior distributions (in the general sense of priors constructed using formal rules) for the model parameters, specifically a Jeffreys prior, the maximal data information (MDI) prior and independent uniform priors on separate model parameters. We investigate the important issue of whether these improper priors lead to proper posterior distributions. We show that, in the GP and GEV cases, the MDI prior, unless modified, never yields a proper posterior and that in the GEV case this also applies to the Jeffreys prior. We also show that a sample size of three (four) is sufficient for independent uniform priors to yield a proper posterior distribution in the GP (GEV) case.

Multiply imputing missing values in data sets with mixed measurement scales using a sequence of generalised linear models

Multiple imputation is a commonly used approach to deal with missing values. In this approach, an imputer repeatedly imputes the missing values by taking draws from the posterior predictive distribution for the missing values conditional on the observed values, and releases these completed data sets to analysts. With each completed data set the analyst performs the analysis of interest, treating the data as if it were fully observed. These analyses are then combined with standard combining rules, allowing the analyst to make appropriate inferences which take into account the uncertainty present due to the missing data. In order to preserve the statistical properties present in the data, the imputer must use a plausible distribution to generate the imputed values. In data sets containing variables with different measurement scales, e.g. some categorical and some continuous variables, this is a challenging problem. A method is proposed to multiply impute missing values in such data sets by modelling the joint distribution of the variables in the data through a sequence of generalised linear models, and data augmentation methods are used to draw imputations from a proper posterior distribution using Markov Chain Monte Carlo (MCMC). The performance of the proposed method is illustrated using simulation studies and on a data set taken from a breast feeding study.

Bayesian Regularization via Graph Laplacian

Regularization plays a critical role in modern statistical research, especially in high-dimensional variable selection problems. Existing Bayesian methods usually assume independence between variables a priori. In this article, we propose a novel Bayesian approach, which explicitly models the dependence structure through a graph Laplacian matrix. We also generalize the graph Laplacian to allow both positively and negatively correlated variables. A prior distribution for the graph Laplacian is then proposed, which allows conjugacy and thereby greatly simplifies the computation. We show that the proposed Bayesian model leads to proper posterior distribution. Connection is made between our method and some existing regularization methods, such as Elastic Net, Lasso, Octagonal Shrinkage and Clustering Algorithm for Regression (OSCAR) and Ridge regression. An efficient Markov Chain Monte Carlo method based on parameter augmentation is developed for posterior computation. Finally, we demonstrate the method through several simulation studies and an application on a real data set involving key performance indicators of electronics companies.

자동차보험 신뢰도 적용에 대한 베이지안 추론 방식 연구

본 연구는 가격 경쟁으로 인하여 최근 들어 요율 세분화가 심화되고 있는 자동차보험 시장에서, 부분 신뢰도의 적용 대상에 대한 경험적 사전분포 (empirical prior distribution) 함수 또는 무정보적 사전분포 (noninformative prior distribution) 정보의 가정을 통한 신뢰도 산출 방식에 대하여 살펴보았다. 요율 세분화의 확대로 가격 산출 단위의 수가 증가될 경우, 부분 신뢰도의 적용 대상은 점차 증가되게 될 것으로 판단되기 때문에, 기존에 제시된 신뢰도 적용 방식을 베이지안 프레임에서 적용, 추론함으로써 보다 다양하고 정교한 방식으로 그 활용 범위를 넓히고자 한다. 즉, 경험적으로 사용되는 사전 분포함수 또는 무정보적 사전 정보를 통하여 적절한 사후분포 (posterior distribution)함수를 도출하고 오차를 최소화하는 베이즈 통계량을 적용한 신뢰도를 추정하여 적용함으로써, 위험도 예측에 있어 기존에 제시된 방법과 비교하여 그 효용성을 입증하고자 한다. 현재 가장 많이 활용되는 제곱근 법칙 (square root rule)의 신뢰도 추정 방식에 베이지안 추론에서 도출된 통계량을 반영한 결과를 분석하여 실질적인 위험도에 수렴하는 수준을 비교하게 된다. 이는 이론적으로 위험도 예측에서 오차를 줄이는 방식에 대한 대안 제시와 더불어 신뢰도 적용 방식에 대한 추가적인 활용 대안을 보험업계에 제시함으로써 요율 세분화로 인한 부분 신뢰도 적용방식에 대한 그 이해와 활용의 폭을 넓히고자 한다. This paper studies the partial credibility application method by assuming the empirical prior or noninformative prior informations in auto insurnace business where intensive rating segmentation is expanded because of premium competition. Expanding of rating factor segmetation brings the increase of pricing cells, as a result, the number of cells for partial credibility application will increase correspondingly. This study is trying to suggest more accurate estimation method by considering the Bayesian framework. By using empirically well-known or noninformative information, inducing the proper posterior distribution and applying the Bayes estimate which is minimizing the error loss into the credibility method, we will show the advantage of Bayesian inference by comparison with current approaches. The comparison is implemented with square root rule which is a widely accepted method in insurance business. The convergence level towarding to the true risk will be compared among various approaches. This study introduces the alternative way of redcuing the error to the auto insurance business fields in need of various methods because of more segmentations.

Objective Bayesian Analysis of a Measurement Error Small Area Model

We consider small area estimation under a nested error linear regression model with measurement errors in the covariates. We propose an objective Bayesian analysis of the model to estimate the finite population means of the small areas. In particular, we derive Jeffreys’ prior for model parameters. We also show that Jeffreys’ prior, which is improper, leads, under very general conditions, to a proper posterior distribution. We have also performed a simulation study where we have compared the Bayes estimates of the finite population means under the Jeffreys’ prior with other Bayesian estimates obtained via the use of the standard flat prior and with non-Bayesian estimates, i.e., the corresponding empirical Bayes estimates and the direct estimates.

Objective Bayesian analysis for exponential power regression models

We develop objective Bayesian analysis for the linear regression model with random errors distributed according to the exponential power distribution. More specifically, we derive explicit expressions for three different Jeffreys priors for the model parameters. We show that only one of these Jeffreys priors leads to a proper posterior distribution. In addition, we develop fast posterior analysis based on Laplace approximations. Moreover, we show that our proposed Bayesian analysis compares favorably to a posterior analysis based on a competing noninformative prior. Finally, we illustrate our methodology with applications of the exponential power regression model to two different datasets.

A Matching Prior for the Shape Parameter of the Skew‐Normal Distribution

Abstract. This paper deals with the issue of performing a default Bayesian analysis on the shape parameter of the skew‐normal distribution. Our approach is based on a suitable pseudo‐likelihood function and a matching prior distribution for this parameter, when location (or regression) and scale parameters are unknown. This approach is important for both theoretical and practical reasons. From a theoretical perspective, it is shown that the proposed matching prior is proper thus inducing a proper posterior distribution for the shape parameter, also when the likelihood is monotone. From the practical perspective, the proposed approach has the advantages of avoiding the elicitation on the nuisance parameters and the computation of multidimensional integrals.

Forecasting volatility with asymmetric smooth transition dynamic range models

Abstract We propose a nonlinear smooth transition conditional autoregressive range (CARR) model for capturing smooth volatility asymmetries in international financial stock markets, building on recent work on smooth transition conditional duration modelling. An adaptive Markov chain Monte Carlo scheme is developed for Bayesian estimation, volatility forecasting and model comparison for the proposed model. The model can capture sign or size asymmetry and heteroskedasticity, such as that which is commonly observed in financial markets. A mixture proposal distribution is developed in order to improve the acceptance rate and the mixing issues which are common in random walk Metropolis-Hastings methods. Further, the logistic transition function is employed and its main properties are considered and discussed in the context of the proposed model, which motivates a suitable, weakly informative prior which ensures a proper posterior distribution and identification of the estimators. The methods are illustrated using simulated data, and an empirical study also provides evidence in favour of the proposed model when forecasting the volatility in two financial stock markets. In addition, the deviance information criterion is employed to compare the proposed models with their limiting classes, the nonlinear threshold CARR models and the symmetric CARR model.

A weighted strategy to handle likelihood uncertainty in Bayesian inference

The sensitivity of posterior inferences to model specification can be considered as an indicator of the presence of outliers, that are to be considered as highly unlikely values under the assumed model. The occurrence of anomalous values can seriously alter the shape of the likelihood function and lead to posterior distributions far from those one would obtain without these data inadequacies. In order to deal with these hindrances, a robust approach is discussed, which allows us to obtain outliers' resistant posterior distributions with properties similar to those of a proper posterior distribution. The methodology is based on the replacement of the genuine likelihood by a weighted likelihood function in the Bayes' formula.

Equality and inequality constrained multivariate linear models: Objective model selection using constrained posterior priors

In objective Bayesian model selection, a well-known problem is that standard non-informative prior distributions cannot be used to obtain a sensible outcome of the Bayes factor because these priors are improper. The use of a small part of the data, i.e., a training sample, to obtain a proper posterior prior distribution has become a popular method to resolve this issue and seems to result in reasonable outcomes of default Bayes factors, such as the intrinsic Bayes factor or a Bayes factor based on the empirical expected-posterior prior. In this paper, it will be illustrated that such default methods may not result in sensible outcomes when evaluating inequality constrained models that are supported by the data. To resolve this issue, a default method is proposed for constructing so-called constrained posterior priors, which are inspired by the symmetrical intrinsic priors discussed by Berger and Mortera (1999) for a simple inequality constrained model selection problem. The resulting Bayes factors can be called “balanced” because model complexity of inequality constrained models is incorporated according to a specific definition that is presented in this paper.

Bayesian analysis of elapsed times in continuous‐time Markov chains

AbstractThe authors consider Bayesian analysis for continuous‐time Markov chain models based on a conditional reference prior. For such models, inference of the elapsed time between chain observations depends heavily on the rate of decay of the prior as the elapsed time increases. Moreover, improper priors on the elapsed time may lead to improper posterior distributions. In addition, an infinitesimal rate matrix also characterizes this class of models. Experts often have good prior knowledge about the parameters of this matrix. The authors show that the use of a proper prior for the rate matrix parameters together with the conditional reference prior for the elapsed time yields a proper posterior distribution. The authors also demonstrate that, when compared to analyses based on priors previously proposed in the literature, a Bayesian analysis on the elapsed time based on the conditional reference prior possesses better frequentist properties. The type of prior thus represents a better default prior choice for estimation software.