In many scientific disciplines complex computer models are used to understand the behaviour of large scale physical systems. An uncertainty analysis of such a computer model known as Galform is presented. Galform models the creation and evolution of approximately one million galaxies from the beginning of the Universe until the current day, and is regarded as a state-of-the-art model within the cosmology community. It requires the specification of many input parameters in order to run the simulation, takes significant time to run, and provides various outputs that can be compared with real world data. A Bayes Linear approach is presented in order to identify the subset of the input space that could give rise to acceptable matches between model output and measured data. This approach takes account of the major sources of uncertainty in a consistent and unified manner, including input parameter uncertainty, function uncertainty, observational error, forcing function uncertainty and structural uncertainty. The approach is known as History Matching, and involves the use of an iterative succession of emulators (stochastic belief specifications detailing beliefs about the Galform function), which are used to cut down the input parameter space. The analysis was successful in producing a large collection of model evaluations that exhibit good fits to the observed data.

This paper develops particle learning (PL) methods for the estimation of general mixture models. The approach is distinguished from alternative particle filtering methods in two major ways. First, each iteration begins by resampling particles according to posterior predictive probability, leading to a more efficient set for propagation. Second, each particle tracks only the "essential state vector" thus leading to reduced dimensional inference. In addition, we describe how the approach will apply to more general mixture models of current interest in the literature; it is hoped that this will inspire a greater number of researchers to adopt sequential Monte Carlo methods for fitting their sophisticated mixture based models. Finally, we show that PL leads to straightforward tools for marginal likelihood calculation and posterior cluster allocation.

The Edgeworth expansion is a series that approximates a probability distribution in terms of its cumulants. One can derive it by first expanding the probability distribution in Hermite orthogonal functions and then collecting terms in powers of the sample size. This paper derives an expansion for posterior distributions which possesses these features of an Edgeworth series. The techniques used are a version of Stein's Identity and properties of Hermite polynomials. Two examples are provided to illustrate the accuracy of our series.

Objective priors have been used in Bayesian models for estimating the number of species in a population, but they have not been examined in depth. Here we derive the form of two objective priors, using Bernardo's reference method and Jeffreys' rule, based on the mixed-Poisson likelihood used in the single-abundance-sample species problem. These derivations are based on asymptotic results for estimates of integer-valued parameters. The factored form of these priors justifies the use of independent prior distributions for the parameter of interest (the number of species) and the nuisance parameters (of the stochastic abundance distribution). We find that the reference prior is preferable overall to the prior resulting from Jeffreys' rule. Although a comprehensive objective Bayesian approach can become analytically intractable for more complicated models, the essence of the approach can be upheld in practice. We analyze several datasets to show that the method can be implemented in practice and that it yields good results, comparable with current competing methods.

We explore the use of importance sampling based on signed root log-likelihood ratios for Bayesian computation. Approximations based on signed root log-likelihood ratios are used in two distinct ways; firstly, to define an importance function and, secondly, to define suitable control variates for variance reduction. These considerations give rise to alternative simulation-consistent schemes to MCMC for Bayesian computation in moderately parameterized regular problems. The schemes based on control variates can also be viewed as usefully supplementing computations based on asymptotic approximations by supplying external estimates of error. The methods are illustrated by a genetic linkage model and a censored regression model.

We consider the problem of analyzing the heterogeneity of clustering distributions for multiple groups of observed data, each of which is indexed by a covariate value, and inferring global clusters arising from observations aggregated over the covariate domain. We propose a novel Bayesian nonparametric method reposing on the formalism of spatial modeling and a nested hierarchy of Dirichlet processes. We provide an analysis of the model properties, relating and contrasting the notions of local and global clusters. We also provide an efficient inference algorithm, and demonstrate the utility of our method in several data examples, including the problem of object tracking and a global clustering analysis of functional data where the functional identity information is not available.