Search

International audience; The formulation of variable selection has been widely developed in the Bayesian literature by linking a random binary indicator to each variable. This Bayesian inference has the advantage of stochastically exploring the set of possible sub-models, whatever their dimension. Bayesian selection approaches, appropriate for categorical predictors, are generally beyond the scope of the standard Bayesian selection of regressors in the linear model since all levels of a categorical variable should be jointly handled in the selection procedure. For categorical covariates, new strategies have been developed to detect the effect of grouped covariates rather than the single effect of a quantitative regressor. In this paper, we review three Bayesian selection methods for categorical predictors: Bayesian Group Lasso with Spike and Slab priors, Bayesian Sparse Group Selection and Bayesian Effect Fusion using model-based clustering. The motivation behind this paper is to provide detailed information about the implementation of the three Bayesian selection methods mentioned above, appropriate for categorical predictors, using the JAGS software. Selection performance and sensitivity analysis of the hyperparameters tuning for prior specifications are assessed under various simulated scenarios. JAGS helps user implement these three Bayesian selection methods for more complex model structures such as hierarchical ones with latent layers.

AMS subject classifications. 62K99, 62L05, 60G15; Making good predictions of a physical system using a computer code requires the inputs to be carefully specified. Some of these inputs, called control variables, reproduce physical conditions, whereas other inputs, called parameters, are specific to the computer code and most often uncertain. The goal of statistical calibration consists in reducing their uncertainty with the help of a statistical model which links the code outputs with the field measurements. In a Bayesian setting, the posterior distribution of these parameters is typically sampled using Markov Chain Monte Carlo methods. However, they are impractical when the code runs are highly time-consuming. A way to circumvent this issue consists of replacing the computer code with a Gaussian process emulator, then sampling a surrogate posterior distribution based on it. Doing so, calibration is subject to an error which strongly depends on the numerical design of experiments used to fit the emulator. Under the assumption that there is no code discrepancy, we aim to reduce this error by constructing a sequential design by means of the expected improvement criterion. Numerical illustrations in several dimensions assess the efficiency of such sequential strategies.

Meteorological ensemble members are a collection of scenarios for future weather issued by a meteorological center. Such ensembles nowadays form the main source of valuable information for probabilistic forecasting which aims at producing a predictive probability distribution of the quantity of interest instead of a single best guess point-wise estimate. Unfortunately, ensemble members cannot generally be considered as a sample from such a predictive probability distribution without a preliminary post-processing treatment to re-calibrate the ensemble. Two main families of post-processing methods, either competing such as the BMA or collaborative such as the EMOS, can be found in the literature. This paper proposes a mixed-effect model belonging to the collaborative family. The structure of the model is formally justified by Bruno de Finetti’s representation theorem which shows how to construct operational statistical models of ensemble based on judgments of invariance under the relabeling of the members. Its interesting specificities are as follows: (1) exchangeability contributes to parsimony, with an interpretation of the latent pivot of the ensemble in terms of a statistical synthesis of the essential meteorological features of the ensemble members, (2) a multiensemble implementation is straightforward, allowing to take advantage of various information so as to increase the sharpness of the forecasting procedure. Focus is cast onto normal statistical structures, first with a direct application for temperatures, then with its very convenient Tobit extension for precipitation. Inference is performed by expectation maximization (EM) algorithms with both steps leading to explicit analytic expressions in the Gaussian temperature case, and recourse is made to stochastic conditional simulations in the zero-inflated precipitation case. After checking its good behavior on artificial data, the proposed post-processing technique is applied to temperature and precipitation ensemble forecasts produced for lead times from 1 to 9 days over five river basins managed by Hydro-Quebec, which ranks among the world’s largest electric companies. These ensemble forecasts, provided by three meteorological global forecast centers (Canadian, USA and European), were extracted from the THORPEX Interactive Grand Global Ensemble (TIGGE) database. The results indicate that post-processed ensembles are calibrated and generally sharper than the raw ensembles for the five watersheds under study. Supplementary materials accompanying this paper appear on-line.

[Departement_IRSTEA]Eaux [TR1_IRSTEA]RIVAGE; In this paper, we propose a novel approach for generating avalanche hazard maps based on the spatial dependence of avalanche runout altitudes. The right-truncated data are described with a Bayesian hierarchical model in which the spatio-temporal process is assumed to be the sum of independent spatial and temporal terms. Topography is roughly taken into account according to valley altitude and path exposition, and the spatial dependence is modelled with a Matérn covariance function. An application is performed to the Haute-Savoie region, French Alps. A spatial dependence in runout altitudes is identified, and an effective range of about 10 km is inferred. The temporal trend extracted highlights the increase of avalanche runout altitudes from 1955, attributed to both anthropogenic factors and climate warming. In a cross validation scheme, spatial predictions are provided on undocumented paths using kriging equations. All in all, although our model is unable to take into account small topographic features, it is a first-ever approach that produces very encouraging results. It could be enhanced in future work by incorporating a numerical physically-based code into the modelling.

Although modeling complex environmental spatio-temporal processes, separability between time and space effects is often assumed as a first approach, even though it may be unrealistic. Hence, the development of adapted non-separable spatio-temporal models is a challenging issue. In this work, we are interested in modeling avalanche counts over the whole French Alps, a large area that exhibits zones characterized by distinct climatic behaviors. We have at our disposal a data set of 405, 63-year series of avalanche counts, one series for one township. We assume the series can be divided into several spatial clusters presenting specific temporal trends. On the latent layer of a hierarchical negative binomial-lognormal structure, the clusters are characterized by independent temporal evolutions, modeled using smoothing splines. The clusters arise as the result of a multinomial probit regression, with spatial coordinates used as covariates. Inference is performed within the Bayesian hierarchical framework. The number of clusters is assessed using cross validation. The best score is obtained for a three-cluster model, but other configurations also share some evidence. The first cluster is located in the east of the French Alps, whereas the others are located together in the west part of the French Alps. The three contrasted temporal trends corresponding to the three clusters result from regional climate change effects, interacting with a strong altitudinal control and data limitation. The proposed model could be used for analyzing various other climatic and environmental data sets. Copyright (c) 2012 John Wiley & Sons, Ltd.

Assessing within-batch and between-batch variability is of major interest for risk assessors and risk managers in the context of microbiological contamination of food. For example, the ratio between the within-batch variability and the between-batch variability has a large impact on the results of a sampling plan. Here, we designed hierarchical Bayesian models to represent such variability. Compatible priors were built mathematically to obtain sound model comparisons. A numeric criterion is proposed to assess the contamination structure comparing the ability of the models to replicate grouped data at the batch level using a posterior predictive loss approach. Models were applied to two case studies: contamination by Listeria monocytogenes of pork breast used to produce diced bacon and contamination by the same microorganism on cold smoked salmon at the end of the process. In the first case study, a contamination structure clearly exists and is located at the batch level, that is, between batches variability is relatively strong, whereas in the second a structure also exists but is less marked.

Complex physical systems are increasingly modeled by computer codes which aim at predicting the reality as accurately as possible. During the last decade, code validation has benefited from a large interest within the scientific community because of the requirement to assess the uncertainty affecting the code outputs. Inspiring frompast contributions to this task, a testing procedure is proposed in this paper to decide either a pure code prediction or a discrepancy-corrected one should be used to provide the best approximation of the physical system. In a particular case where the computer code depends on uncertain parameters, this problem of model selection can be carried out in a Bayesian setting. It requires the specification of proper prior distributions that are well known as having a strong impact on the results. Another way consists in specifying non-informative priors. However, they are sometimes improper, which is a major barrier for computing the Bayes factor. A way to overcome this issue is to use the so-called intrinsic Bayes factor (IBF) in order to replace the ill-defined Bayes factor when improper priors are used. For computer codes which depend linearly on their parameters, the computation of the IBF is made easier, thanks to some explicit marginalization. In the paper, we present a special case where the IBF is equal to the standard Bayes factor when the right-Haar prior is specified on the code parameters and the scale of the code discrepancy. On simulated data, the IBF has been computed for several prior distributions. A confounding effect between the code discrepancy and the linear code is pointed out. Finally, the IBF is computed for an industrial computer code used for monitoring power plant production.

In the context of risk analysis under uncertainty, we focus here on the problem of estimating a so-called quantity of interest of an uncertainty analysis problem, i.e. a given feature of the probability distribution function (pdf) of the output of a deterministic model with uncertain inputs. We will stay here in a fully probabilistic setting. A common problem is how to account for epistemic uncertainty tainting the parameter of the probability distribution of the inputs. In the standard practice, this uncertainty is often neglected (plug-in approach). When a specific uncertainty assessment is made, under the basis of the available information (expertise and/or data), a common solution consists in marginalizing the joint distribution of both observable inputs and parameters of the probabilistic model (i.e. computing the predictive pdf of the inputs), then propagating it through the deterministic model. We will reinterpret this approach in the light of Bayesian decision theory, and will put into evidence that this practice leads the analyst to adopt implicitly a specific loss function which may be inappropriate for the problem under investigation, and suboptimal from a decisional perspective. These concepts are illustrated on a simple numerical example, concerning a case of flood risk assessment.

Hierarchical Bayesian Modelling is powerful however under-used to model and evaluate the risks associated with the development of pathogens in food industry, to predict exotic invasions, species extinctions and development of emerging diseases, or to assess chemical risks. Modelling population dynamics of Saccharomyces cerevisiae considering its biodiversity and other sources of variability is crucial for selecting strains meeting industrial needs. Using this approach, we studied the population dynamics of S. cerevisiae, the domesticated yeast. widely encountered in food industry, notably in brewery, vinery, bakery and distillery. We relied on a logistic equation to estimate the key variables of population growth, but we took also into account factors able to affect them, namely environmental effects, genetic diversity and measurement errors. Our probabilistic approach allowed us: (i) to model the dynamical behaviour of strains in a given condition under some uncertainty, (ii) to measure environmental effects and (iii) to evaluate genetic variability of the growth key variables.

A parsimonious model is presented as an alternative to delta approaches to modelling zero-inflated continuous data. The data model relies on an exponentially compound Poisson process, also called the law of leaks (LOL). It represents the process of sampling resources that are spatially distributed as Poisson distributed patches, each containing a certain quantity of biomass drawn from an exponential distribution. In an application of the LOL, two latent structures are proposed to account for spatial dependencies between zero values at different scales within a hierarchical Bayesian framework. The LOL is compared to the delta-gamma (Delta I") distribution using bottom-trawl survey data. Results of this case study emphasize that the LOL provides slightly better fits to learning samples with a very high proportion of zero values and small strictly positive abundance data. Additionally, it offers better predictions of validation samples.