Multi-source data fusion and super-resolution from astronomical images

Virtual observatories give us access to huge amounts of image data that are often redundant. Our goal is to take advantage of this redundancy by combining images of the same field of view into a single model. To achieve this goal, we propose to develop a multi-source data fusion method that relies on probability and band-limited signal theory. The target object is an image to be inferred from a number of blurred and noisy sources, possibly from different sensors under various conditions (i.e. resolution, shift, orientation, blur, noise...). We aim at the recovery of a compound model “image + uncertainties” that best relates to the observations and contains a maximum of useful information from the initial data set. Thus, in some cases, spatial super-resolution may be required in order to preserve the information. We propose to use a Bayesian inference scheme to invert a forward model, which describes the image formation process for each observation and takes into account some a priori knowledge (e.g. stars as point sources). This involves both automatic registration and spatial resampling, which are ill-posed inverse problems that are addressed within a rigorous Bayesian framework. The originality of the work is in devising a new technique of multi-image data fusion that provides us with super-resolution, self-calibration and possibly model selection capabilities. This approach should outperform existing methods such as resample-and-add or drizzling since it can handle different instrument characteristics for each input image and compute uncertainty estimates as well. Moreover, it is designed to also work in a recursive way, so that the model can be updated when new data become available.     

Semiparametric smoothing of sparse contingency tables

In the paper simple resampling technique based on semiparametric smoothing is introduced. Although the method is very flexible and in principle can be applied to any sparse data and ill-posed statistical problem, its efficient or even reasonable implementation requires special investigation. In the paper a problem of fitting local dependence structure of finite-state random sequences is addressed. This problem is relevant, for example, in genetics, bioinformatics, computer linguistics, etc., and usually leads to analysis of sparse contingency tables of dependent categorical data. Thus, the classical assumptions of log-linear model, a standard technique for analysis of contingency tables, do not hold. A framework convenient for implementation of semiparametric smoothing and resampling is proposed. It is based on a special representation form of data under consideration and generalized logit model. A computer experiment is carried out to gain better insight on practical performance of the procedure.     

A bayesian alternative to the analysis of matched categorical responces

This study investigates the Bayesian appeoach to the analysis of parired responess when the responses are categorical. Using resampling and analytical procedures, inferences for homogeneity and agreement are develped. The posterior analysis is based on the Dirichlet distribution from which repeated samples can be geneated with a random number generator. Resampling and analytical techniques are employed to make Bayesian inferences, and when it is not appropriate to use analytical procedures, resampling techniques are easily implemented. Bayesian methodoloogy is illustrated with several examples and the results show that they are exacr-small sample procedures that can easily solve inference problems for matched designs.     

MULTIPLE COMPARISONS OF LOG-LIKELIHOODS AND COMBINING NONNESTED MODELS WITH APPLICATIONS TO PHYLOGENETIC TREE SELECTION

We consider multiple comparisons of log-likelihood's to take account of the multiplicity of testings in selection of nonnested models. A resampling version of the Gupta procedure for the selection problem is used to obtain a set of good models, which are not significantly worse than the maximum likelihood model; i.e., a confidence set of models. Our method is to test which model is better than the other, while the object of the classical testing methods is to find the correct model. Thus the null hypotheses behind these two approaches are very different. Our method and the other commonly used approaches, such as the approximate Bayesian posterior, the bootstrap selection probability, and the LR test against the full model, are applied to the selection of molecular phylogenetic tree of mammal species. Tree selection is a version of the model-based clustering, which is an example of nonnested model selection. It is shown that the structure of the tree selection problem is equivalent to that of the variable selection problem of the multiple regression with some constraints on the combinations of the variables. It turns out that the LR test rejects all the possible trees because of the misspecification of the models, whereas our method gives a reasonable confidence set. For a better understanding of the uncertainty in the selection, we combine the maximum likelihood estimates (MLE's) of the trees to obtain the full model that includes the trees as the submodels by using a linear approximation of the parametric models. The MLE of the phylogeny is then represented as a network of species rather than a tree. A geometrical interpretation of the problem is also discussed.     

Partially linear transformation model for length-biased and right-censored data

In this paper, we consider a partially linear transformation model for data subject to length-biasedness and right-censoring which frequently arise simultaneously in biometrics and other fields. The partially linear transformation model can account for nonlinear covariate effects in addition to linear effects on survival time, and thus reconciles a major disadvantage of the popular semiparamnetric linear transformation model. We adopt local linear fitting technique and develop an unbiased global and local estimating equations approach for the estimation of unknown covariate effects. We provide an asymptotic justification for the proposed procedure, and develop an iterative computational algorithm for its practical implementation, and a bootstrap resampling procedure for estimating the standard errors of the estimator. A simulation study shows that the proposed method performs well in finite samples, and the proposed estimator is applied to analyse the Oscar data.     

Bayesian,classical and hybrid methods of inference when one parameter value is special

This paper considers the problem of making statistical inferences about a parameter when a narrow interval centred at a given value of the parameter is considered special, which is interpreted as meaning that there is a substantial degree of prior belief that the true value of the parameter lies in this interval. A clear justification of the practical importance of this problem is provided. The main difficulty with the standard Bayesian solution to this problem is discussed and, as a result, a pseudo-Bayesian solution is put forward based on determining lower limits for the posterior probability of the parameter lying in the special interval by means of a sensitivity analysis. Since it is not assumed that prior beliefs necessarily need to be expressed in terms of prior probabilities, nor that post-data probabilities must be Bayesian posterior probabilities, hybrid methods of inference are also proposed that are based on specific ways of measuring and interpreting the classical concept of significance. The various methods that are outlined are compared and contrasted at both a foundational level, and from a practical viewpoint by applying them to real data from meta-analyses that appeared in a well-known medical article.     

Objective Testing Procedures in Linear Models: Calibration of the p-values

Abstract.  An optimal Bayesian decision procedure for testing hypothesis in normal linear models based on intrinsic model posterior probabilities is considered. It is proven that these posterior probabilities are simple functions of the classical F -statistic, thus the evaluation of the procedure can be carried out analytically through the frequentist analysis of the posterior probability of the null. An asymptotic analysis proves that, under mild conditions on the design matrix, the procedure is consistent. For any testing hypothesis it is also seen that there is a one-to-one mapping – which we call calibration curve – between the posterior probability of the null hypothesis and the classical bi p -value. This curve adds substantial knowledge about the possible discrepancies between the Bayesian and the p -value measures of evidence for testing hypothesis. It permits a better understanding of the serious difficulties that are encountered in linear models for interpreting the p -values. A specific illustration of the variable selection problem is given.     

Bayesian estimation of traffic intensity based on queue length in a multi-server M/M/s queue

In this article, we focus on multi-server queueing systems in which inter-arrival and service times are exponentially distributed (Markovian). We use a Bayesian technique, the sampling/importance resampling method (SIR), to estimate the parameters of these queueing systems, making possible the determination of performance measures that are essential to the evaluation of important practical applications such as computer and telecommunication networks, manufacturing and service systems, health care, and other similar real-life problems. Extensive numerical results are presented to demonstrate the accuracy and efficiency of the technique, as well as some of its limitations.     

A BAYESIAN ANALYSIS FOR INTER-RATER AGREEMENT

An analysis of inter-rater agreement is presented. We study the problem with several raters using a Bayesian model based on the Dirichlet distribution. Inter-rater agreement, including global and partial agreement, is studied by determining the joint posterior distribution of the raters. Posterior distributions are computed with a direct resampling technique. Our method is illustrated with an example involving four residents, who are diagnosing 12 psychiatric patients suspected of having a thought disorder. Initially employing analytical and resampling methods, total agreement between the four is examined with a Bayesian testing technique. Later, partial agreement is examined by determining the posterior probability of certain orderings among the rater means. The power of resampling is revealed by its ability to compute complex multiple integrals that represent various posterior probabilities of agreement and disagreement between several raters.     

Bayesian D-Optimal Designs for Poisson Regression Models

By incorporating informative and/or historical knowledge of the unknown parameters, Bayesian experimental design under the decision-theory framework can combine all the information available to the experimenter so that a better design may be achieved. Bayesian optimal designs for generalized linear regression models, especially for the Poisson regression model, is of interest in this article. In addition, lack of an efficient computational method in dealing with the Bayesian design leads to development of a hybrid computational method that consists of the combination of a rough global optima search and a more precise local optima search. This approach can efficiently search for the optimal design for multi-variable generalized linear models. Furthermore, the equivalence theorem is used to verify whether the design is optimal or not.     

Microdata disclosure by resampling – Empirical findings for business survey data

Summary: One specific problem statistical offices and research institutes are faced with when releasing microdata is the preservation of confidentiality. Traditional methods to avoid disclosure often destroy the structure of the data, and information loss is potentially high. In this paper an alternative technique of creating scientific–use files is discussed, which reproduces the characteristics of the original data quite well. It is based on Fienberg (1997, 1994) who estimates and resamples from the empirical multivariate cumulative distribution function of the data in order to get synthetic data. The procedure creates data sets – the resample – which have the same characteristics as the original survey data. The paper includes some applications of this method with (a) simulated data and (b) innovation survey data, the Mannheim Innovation Panel (MIP), and a comparison between resampling and a common method of disclosure control (disturbance with multiplicative error) with regard to confidentiality on the one hand and the appropriateness of the disturbed data for different kinds of analyses on the other. The results show that univariate distributions can be better reproduced by unweighted resampling. Parameter estimates can be reproduced quite well if the resampling procedure implements the correlation structure of the original data as a scale or if the data is multiplicatively perturbed and a correction term is used. On average, anonymization of data with multiplicatively perturbed values protects better against re–identification than the various resampling methods used.     

Bayesian Additive Machine: classification with a semiparametric discriminant function

In this paper, we propose a new Bayesian inference approach for classification based on the traditional hinge loss used for classical support vector machines, which we call the Bayesian Additive Machine (BAM). Unlike existing approaches, the new model has a semiparametric discriminant function where some feature effects are nonlinear and others are linear. This separation of features is achieved automatically during model fitting without user pre-specification. Following the literature on sparse regression of high-dimensional models, we can also identify the irrelevant features. By introducing spike-and-slab priors using two sets of indicator variables, these multiple goals are achieved simultaneously and automatically, without any parameter tuning such as cross-validation. An efficient partially collapsed Markov chain Monte Carlo algorithm is developed for posterior exploration based on a data augmentation scheme for the hinge loss. Our simulations and three real data examples demonstrate that the new approach is a strong competitor to some approaches that were proposed recently for dealing with challenging classification examples with high dimensionality.     

A Bayesian analysis of the change-point problem for directional data

In this paper, we discuss a simple fully Bayesian analysis of the change-point problem for the directional data in the parametric framework with von Mises or circular normal distribution as the underlying distribution. We first discuss the problem of detecting change in the mean direction of the circular normal distribution using a latent variable approach when the concentration parameter is unknown. Then, a simpler approach, beginning with proper priors for all the unknown parameters – the sampling importance resampling technique – is used to obtain the posterior marginal distribution of the change-point. The method is illustrated using the wind data [E.P. Weijers, A. Van Delden, H.F. Vugts and A.G.C.A. Meesters, The composite horizontal wind field within convective structures of the atmospheric surface layer, J. Atmos. Sci. 52 (1995. 3866–3878]. The method can be adapted for a variety of situations involving both angular and linear data and can be used with profit in the context of statistical process control in Phase I of control charting and also in Phase II in conjunction with control charts.     

Hierarchical bayesian curve fitting and smoothing

Estimation of a smooth function is considered when observations on this function added with Gaussian errors are observed. The problem is formulated as a general linear model, and a hierarchical Bayesian approach is then used to study it. Credible bands are also developed for the function. Sensitivity analysis is conducted to determine the influence of the choice of priors on hyperparameters. Finally, the methodology is illustrated using real and simulated examples where it is compared with classical cubic splines. It is also shown that our approach provides a Bayesian solution to some problems in discrete time series.     

On a bayesian approach for the shiftpoint problem

Using a direct resampling process, a Bayesian approach is developed for the analysis of the shiftpoint problem. In many problems it is straight forward to isolate the marginal posterior distribution of the shift-point parameter and the conditional distribution of some of the parameters given the shift point and the other remaining parameters. When this is possible, a direct sampling approach is easily implemented whereby standard random number generators can be used to generate samples from the joint posterior distribution of aii the parameters in the model. This technique is illustrated with examples involving one shift for Poisson processes and regression models.     

Bayesian inference over ICA models: application to multibiometric score fusion with quality estimates

Bayesian networks are not well-formulated for continuous variables. The majority of recent works dealing with Bayesian inference are restricted only to special types of continuous variables such as the conditional linear Gaussian model for Gaussian variables. In this context, an exact Bayesian inference algorithm for clusters of continuous variables which may be approximated by independent component analysis models is proposed. The complexity in memory space is linear and the overfitting problem is attenuated, while the inference time is still exponential. Experiments for multibiometric score fusion with quality estimates are conducted, and it is observed that the performances are satisfactory compared to some known fusion techniques.     

Generalised data augmentation and posterior inferences

This paper explores the use of data augmentation in settings beyond the standard Bayesian one. In particular, we show that, after proposing an appropriate generalised data-augmentation principle, it is possible to extend the range of sampling situations in which fiducial methods can be applied by constructing Markov chains whose stationary distributions represent valid posterior inferences on model parameters. Some properties of these chains are presented and a number of open questions are discussed. We also use the approach to draw out connections between classical and Bayesian approaches in some standard settings.     

Quasi-random resampling for the bootstrap

Quasi-random sequences are known to give efficient numerical integration rules in many Bayesian statistical problems where the posterior distribution can be transformed into periodic functions on then-dimensional hypercube. From this idea we develop a quasi-random approach to the generation of resamples used for Monte Carlo approximations to bootstrap estimates of bias, variance and distribution functions. We demonstrate a major difference between quasi-random bootstrap resamples, which are generated by deterministic algorithms and have no true randomness, and the usual pseudo-random bootstrap resamples generated by the classical bootstrap approach. Various quasi-random approaches are considered and are shown via a simulation study to result in approximants that are competitive in terms of efficiency when compared with other bootstrap Monte Carlo procedures such as balanced and antithetic resampling.     

Sparse regression techniques in low-dimensional survival data settings

In high-dimensional data settings, sparse model fits are desired, which can be obtained through shrinkage or boosting techniques. We investigate classical shrinkage techniques such as the lasso, which is theoretically known to be biased, new techniques that address this problem, such as elastic net and SCAD, and boosting technique CoxBoost and extensions of it, which allow to incorporate additional structure. To examine, whether these methods, that are designed for or frequently used in high-dimensional survival data analysis, provide sensible results in low-dimensional data settings as well, we consider the well known GBSG breast cancer data. In detail, we study the bias, stability and sparseness of these model fitting techniques via comparison to the maximum likelihood estimate and resampling, and their prediction performance via prediction error curve estimates.     

Using historical data for Bayesian sample size determination

Summary.  We consider the sample size determination (SSD) problem, which is a basic yet extremely important aspect of experimental design. Specifically, we deal with the Bayesian approach to SSD, which gives researchers the possibility of taking into account pre-experimental information and uncertainty on unknown parameters. At the design stage, this fact offers the advantage of removing or mitigating typical drawbacks of classical methods, which might lead to serious miscalculation of the sample size. In this context, the leading idea is to choose the minimal sample size that guarantees a probabilistic control on the performance of quantities that are derived from the posterior distribution and used for inference on parameters of interest. We are concerned with the use of historical data—i.e. observations from previous similar studies—for SSD. We illustrate how the class of power priors can be fruitfully employed to deal with lack of homogeneity between historical data and observations of the upcoming experiment. This problem, in fact, determines the necessity of discounting prior information and of evaluating the effect of heterogeneity on the optimal sample size. Some of the most popular Bayesian SSD methods are reviewed and their use, in concert with power priors, is illustrated in several medical experimental contexts.