Bayes and MCMC for Undergraduates

Students of statistics should be taught the ideas and methods that are widely used in practice and that will help them understand the world of statistics. Today, this means teaching them about Bayesian methods. In this article, I present ideas on teaching an undergraduate Bayesian course that uses Markov chain Monte Carlo and that can be a second course or, for strong students, a first course in statistics.     

A tutorial on adaptive MCMC

We review adaptive Markov chain Monte Carlo algorithms (MCMC) as a mean to optimise their performance. Using simple toy examples we review their theoretical underpinnings, and in particular show why adaptive MCMC algorithms might fail when some fundamental properties are not satisfied. This leads to guidelines concerning the design of correct algorithms. We then review criteria and the useful framework of stochastic approximation, which allows one to systematically optimise generally used criteria, but also analyse the properties of adaptive MCMC algorithms. We then propose a series of novel adaptive algorithms which prove to be robust and reliable in practice. These algorithms are applied to artificial and high dimensional scenarios, but also to the classic mine disaster dataset inference problem.     

Mode Jumping Proposals in MCMC

Markov chain Monte Carlo algorithms generate samples from a target distribution by simulating a Markov chain. Large flexibility exists in specification of transition matrix of the chain. In practice, however, most algorithms used only allow small changes in the state vector in each iteration. This choice typically causes problems for multi-modal distributions as moves between modes become rare and, in turn, results in slow convergence to the target distribution. In this paper we consider continuous distributions on R ⁿ and specify how optimization for local maxima of the target distribution can be incorporated in the specification of the Markov chain. Thereby, we obtain a chain with frequent jumps between modes. We demonstrate the effectiveness of the approach in three examples. The first considers a simple mixture of bivariate normal distributions, whereas the two last examples consider sampling from posterior distributions based on previously analysed data sets.     

DRAM: Efficient adaptive MCMC

We propose to combine two quite powerful ideas that have recently appeared in the Markov chain Monte Carlo literature: adaptive Metropolis samplers and delayed rejection. The ergodicity of the resulting non-Markovian sampler is proved, and the efficiency of the combination is demonstrated with various examples. We present situations where the combination outperforms the original methods: adaptation clearly enhances efficiency of the delayed rejection algorithm in cases where good proposal distributions are not available. Similarly, delayed rejection provides a systematic remedy when the adaptation process has a slow start.     

Parallelizing MCMC for Bayesian spatiotemporal geostatistical models

When MCMC methods for Bayesian spatiotemporal modeling are applied to large geostatistical problems, challenges arise as a consequence of memory requirements, computing costs, and convergence monitoring. This article describes the parallelization of a reparametrized and marginalized posterior sampling (RAMPS) algorithm, which is carefully designed to generate posterior samples efficiently. The algorithm is implemented using the Parallel Linear Algebra Package (PLAPACK). The scalability of the algorithm is investigated via simulation experiments that are implemented using a cluster with 25 processors. The usefulness of the method is illustrated with an application to sulfur dioxide concentration data from the Air Quality System database of the U.S. Environmental Protection Agency.     

A Hellinger distance approach to MCMC diagnostics

Bayesian analysis often requires the researcher to employ Markov Chain Monte Carlo (MCMC) techniques to draw samples from a posterior distribution which in turn is used to make inferences. Currently, several approaches to determine convergence of the chain as well as sensitivities of the resulting inferences have been developed. This work develops a Hellinger distance approach to MCMC diagnostics. An approximation to the Hellinger distance between two distributions f and g based on sampling is introduced. This approximation is studied via simulation to determine the accuracy. A criterion for using this Hellinger distance for determining chain convergence is proposed as well as a criterion for sensitivity studies. These criteria are illustrated using a dataset concerning the Anguilla australis, an eel native to New Zealand.     

Bayesian model learning based on a parallel MCMC strategy

We introduce a novel Markov chain Monte Carlo algorithm for estimation of posterior probabilities over discrete model spaces. Our learning approach is applicable to families of models for which the marginal likelihood can be analytically calculated, either exactly or approximately, given any fixed structure. It is argued that for certain model neighborhood structures, the ordinary reversible Metropolis-Hastings algorithm does not yield an appropriate solution to the estimation problem. Therefore, we develop an alternative, non-reversible algorithm which can avoid the scaling effect of the neighborhood. To efficiently explore a model space, a finite number of interacting parallel stochastic processes is utilized. Our interaction scheme enables exploration of several local neighborhoods of a model space simultaneously, while it prevents the absorption of any particular process to a relatively inferior state. We illustrate the advantages of our method by an application to a classification model. In particular, we use an extensive bacterial database and compare our results with results obtained by different methods for the same data.     

R package rjmcmc: reversible jump MCMC using post‐processing

The rjmcmc package for R implements the post‐processing reversible jump Markov chain Monte Carlo (MCMC) algorithm of Barker & Link. MCMC output from each of the models is used to estimate posterior model probabilities and Bayes factors. Automatic differentiation is used to simplify implementation. The package is demonstrated on two examples.     

Markov-chain monte carlo: Some practical implications of theoretical results

We review and discuss some recent progress in the theory of Markov-chain Monte Carlo applications, particularly oriented to applications in statistics. We attempt to assess the relevance of this theory for practical applications.     

A Bayesian prediction using the skew Gaussian distribution

A model based on the skew Gaussian distribution is presented to handle skewed spatial data. It extends the results of popular Gaussian process models. Markov chain Monte Carlo techniques are used to generate samples from the posterior distributions of the parameters. Finally, this model is applied in the spatial prediction of weekly rainfall. Cross-validation shows that the predictive performance of our model compares favorably with several kriging variants.     

A comparison of Bayesian model selection based on MCMC with an application to GARCH-type models

This paper presents a comprehensive review and comparison of five computational methods for Bayesian model selection, based on MCMC simulations from posterior model parameter distributions. We apply these methods to a well-known and important class of models in financial time series analysis, namely GARCH and GARCH-t models for conditional return distributions (assuming normal and t-distributions). We compare their performance with the more common maximum likelihood-based model selection for simulated and real market data. All five MCMC methods proved reliable in the simulation study, although differing in their computational demands. Results on simulated data also show that for large degrees of freedom (where the t-distribution becomes more similar to a normal one), Bayesian model selection results in better decisions in favor of the true model than maximum likelihood. Results on market data show the instability of the harmonic mean estimator and reliability of the advanced model selection methods.     

PHASE RANDOMIZATION: A CONVERGENCE DIAGNOSTIC TEST FOR MCMC

Most Markov chain Monte Carlo (MCMC) users address the convergence problem by applying diagnostic tools to the output produced by running their samplers. Potentially useful diagnostics can be borrowed from diverse areas such as time series. One such method is phase randomization. This paper describes this method in the context of MCMC, summarizes its characteristics, and contrasts its performance with those of the more common diagnostic tests for MCMC. It is observed that the new tool contributes information about third‐ and higher‐order cumulant behaviour which is important in characterizing certain forms of nonlinearity and non‐stationarity.     

MCMC方法下最优Copula的估计及选取

针对目前Copula函数在实际中的应用问题,介绍了一种基于马尔科夫链蒙特卡罗方法（MCMC）的Copula函数估计及选取方法,并将该方法与目前常用方法进行系统比较,最后对上证综合指数和深证成分指数进行了实证分析,结果体现了该法的有效性。     

Note on posterior inference for the Bingham distribution

The properties of high-dimensional Bingham distributions have been studied by Kume and Walker (2014 Kume, A., and S. G. Walker. 2014. On the Bingham distribution with large dimension. Journal of Multivariate Analysis 124:345–52.[Crossref], [Web of Science ®] , [Google Scholar]). Fallaize and Kypraios (2016 Fallaize, C. J., and T. Kypraios. 2016. Exact Bayesian inference for the Bingham distribution. Statistics and Computing 26:349–60.[Crossref], [Web of Science ®] , [Google Scholar]) propose the Bayesian inference for the Bingham distribution and they use developments in Bayesian computation for distributions with doubly intractable normalizing constants (Møller et al. 2006 Møller, J., A. N. Pettitt, R. Reeves, and K. K. Berthelsen. 2006. An efficient Markov chain Monte Carlo method for distributions with intractable normalising constants. Biometrika 93 (2):451–458.[Crossref], [Web of Science ®] , [Google Scholar]; Murray, Ghahramani, and MacKay 2006 Murray, I., Z. Ghahramani, and D. J. C. MacKay. 2006. MCMC for doubly intractable distributions. In Proceedings of the 22nd annual conference on uncertainty in artificial intelligence (UAI-06), 359–66. AUAI Press. [Google Scholar]). However, they rely heavily on two Metropolis updates that they need to tune. In this article, we propose instead a model selection with the marginal likelihood.     

Image segmentation using voronoi polygons and MCMC, with application to muscle fibre images

We investigate a Bayesian method for the segmentation of muscle fibre images. The images are reasonably well approximated by a Dirichlet tessellation, and so we use a deformable template model based on Voronoi polygons to represent the segmented image. We consider various prior distributions for the parameters and suggest an appropriate likelihood. Following the Bayesian paradigm, the mathematical form for the posterior distribution is obtained (up to an integrating constant). We introduce a Metropolis-Hastings algorithm and a reversible jump Markov chain Monte Carlo algorithm (RJMCMC) for simulation from the posterior when the number of polygons is fixed or unknown. The particular moves in the RJMCMC algorithm are birth, death and position/colour changes of the point process which determines the location of the polygons. Segmentation of the true image was carried out using the estimated posterior mode and posterior mean. A simulation study is presented which is helpful for tuning the hyperparameters and to assess the accuracy. The algorithms work well on a real image of a muscle fibre cross-section image, and an additional parameter, which models the boundaries of the muscle fibres, is included in the final model.     

Time use of married couples: Bayesian approach

ABSTRACT

The living hours data of individuals' time spent on daily activities are compositional and include many zeros because individuals do not pursue all activities every day. Thus, we should exercise caution in using such data for empirical analyses. The Bayesian method offers several advantages in analyzing compositional data. In this study, we analyze the time allocation of Japanese married couples using the Bayesian model. Based on the Bayes factors, we compare models that consider and do not consider the correlations between married couples' time use data. The model that considers the correlation shows superior performance. We show that the Bayesian method can adequately take into account the correlations of wives' and husbands' living hours, facilitating the calculation of partial effects that their activities' variables have on living hours. The partial effects of the model that considers the correlations between the couples' time use are easily calculated from the posterior results.     

Inference for Lévy‐Driven Stochastic Volatility Models via Adaptive Sequential Monte Carlo

Abstract. We investigate simulation methodology for Bayesian inference in Lévy‐driven stochastic volatility (SV) models. Typically, Bayesian inference from such models is performed using Markov chain Monte Carlo (MCMC); this is often a challenging task. Sequential Monte Carlo (SMC) samplers are methods that can improve over MCMC; however, there are many user‐set parameters to specify. We develop a fully automated SMC algorithm, which substantially improves over the standard MCMC methods in the literature. To illustrate our methodology, we look at a model comprised of a Heston model with an independent, additive, variance gamma process in the returns equation. The driving gamma process can capture the stylized behaviour of many financial time series and a discretized version, fit in a Bayesian manner, has been found to be very useful for modelling equity data. We demonstrate that it is possible to draw exact inference, in the sense of no time‐discretization error, from the Bayesian SV model.     

Learning a multivariate Gaussian mixture model with the reversible jump MCMC algorithm

This paper is a contribution to the methodology of fully Bayesian inference in a multivariate Gaussian mixture model using the reversible jump Markov chain Monte Carlo algorithm. To follow the constraints of preserving the first two moments before and after the split or combine moves, we concentrate on a simplified multivariate Gaussian mixture model, in which the covariance matrices of all components share a common eigenvector matrix. We then propose an approach to the construction of the reversible jump Markov chain Monte Carlo algorithm for this model. Experimental results on several data sets demonstrate the efficacy of our algorithm.     

Hierarchical random effect models for coastal erosion of cliffs in the Holderness coast

Prediction of possible cliff erosion at some future date is fundamental to coastal planning and shoreline management, for example to avoid development in vulnerable areas. Historically, to predict cliff recession rates deterministic methods were used. More recently, recession predictions have been expressed in probabilistic terms. However, to date, only simplistic models have been developed. We consider the cliff erosion along the Holderness Coast. Since 1951 a monitoring program has been started in 118 stations along the coast, providing an invaluable, but often missing, source of information. We build hierarchical random effect models, taking account of the known dynamics of the process and including the missing information.     

On the Value of derivative evaluations and random walk suppression in Markov Chain Monte Carlo algorithms

Two strategies that can potentially improve Markov Chain Monte Carlo algorithms are to use derivative evaluations of the target density, and to suppress random walk behaviour in the chain. The use of one or both of these strategies has been investigated in a few specific applications, but neither is used routinely. We undertake a broader evaluation of these techniques, with a view to assessing their utility for routine use. In addition to comparing different algorithms, we also compare two different ways in which the algorithms can be applied to a multivariate target distribution. Specifically, the univariate version of an algorithm can be applied repeatedly to one-dimensional conditional distributions, or the multivariate version can be applied directly to the target distribution.