arXiv subjects

Reinhard Furrer

Publications and source records attributed to Reinhard Furrer.

At least 37 records · Page 2Linked to original sources

arXiv2020

Bayesian spatial modelling of terrestrial radiation in Switzerland

The geographic variation of terrestrial radiation can be exploited in epidemiological studies of the health effects of protracted low-dose exposure. Various methods have been applied to derive maps of this variation. We aimed to construct a map of terrestrial radiation for Switzerland. We used airborne $γ$-spectrometry measurements to model the ambient dose rates from terrestrial radiation through a Bayesian mixed-effects model and conducted inference using Integrated Nested Laplace Approximation (INLA). We predicted higher levels of ambient dose rates in the alpine regions and Ticino compared with the western and northern parts of Switzerland. We provide a map that can be used for exposure assessment in epidemiological studies and as a baseline map for assessing potential contamination.

stat.AP↗

arXiv2020

Asymptotically Equivalent Prediction in Multivariate Geostatistics

Cokriging is the common method of spatial interpolation (best linear unbiased prediction) in multivariate geostatistics. While best linear prediction has been well understood in univariate spatial statistics, the literature for the multivariate case has been elusive so far. The new challenges provided by modern spatial datasets, being typically multivariate, call for a deeper study of cokriging. In particular, we deal with the problem of misspecified cokriging prediction within the framework of fixed domain asymptotics. Specifically, we provide conditions for equivalence of measures associated with multivariate Gaussian random fields, with index set in a compact set of a d-dimensional Euclidean space. Such conditions have been elusive for over about 50 years of spatial statistics. We then focus on the multivariate Matérn and Generalized Wendland classes of matrix valued covariance functions, that have been very popular for having parameters that are crucial to spatial interpolation, and that control the mean square differentiability of the associated Gaussian process. We provide sufficient conditions, for equivalence of Gaussian measures, relying on the covariance parameters of these two classes. This enables to identify the parameters that are crucial to asymptotically equivalent interpolation in multivariate geostatistics. Our findings are then illustrated through simulation studies.

math.ST↗

arXiv2020

Multiresolution Decomposition of Areal Count Data

Multiresolution decomposition is commonly understood as a procedure to capture scale-dependent features in random signals. Such methods were first established for image processing and typically rely on raster or regularly gridded data. In this article, we extend a particular multiresolution decomposition procedure to areal count data, i.e.~discrete irregularly gridded data. More specifically, we incorporate in a new model concept and distributions from the so-called Besag--York--Mollié model to include a priori demographical knowledge. These adaptions and subsequent changes in the computation schemes are carefully outlined below, whereas the main idea of the original multiresolution decomposition remains. Finally, we show the extension's feasibility by applying it to oral cavity cancer counts in Germany.

stat.ME↗

arXiv2019

Asymptotic properties of the maximum likelihood and cross validation estimators for transformed Gaussian processes

The asymptotic analysis of covariance parameter estimation of Gaussian processes has been subject to intensive investigation. However, this asymptotic analysis is very scarce for non-Gaussian processes. In this paper, we study a class of non-Gaussian processes obtained by regular non-linear transformations of Gaussian processes. We provide the increasing-domain asymptotic properties of the (Gaussian) maximum likelihood and cross validation estimators of the covariance parameters of a non-Gaussian process of this class. We show that these estimators are consistent and asymptotically normal, although they are defined as if the process was Gaussian. They do not need to model or estimate the non-linear transformation. Our results can thus be interpreted as a robustness of (Gaussian) maximum likelihood and cross validation towards non-Gaussianity. Our proofs rely on two technical results that are of independent interest for the increasing-domain asymptotic literature of spatial processes. First, we show that, under mild assumptions, coefficients of inverses of large covariance matrices decay at an inverse polynomial rate as a function of the corresponding observation location distances. Second, we provide a general central limit theorem for quadratic forms obtained from transformed Gaussian processes. Finally, our asymptotic results are illustrated by numerical simulations.

math.ST↗

arXiv2019

Additive Bayesian Network Modelling with the R Package abn

The R package abn is designed to fit additive Bayesian models to observational datasets. It contains routines to score Bayesian networks based on Bayesian or information theoretic formulations of generalized linear models. It is equipped with exact search and greedy search algorithms to select the best network. It supports a possible blend of continuous, discrete and count data and input of prior knowledge at a structural level. The Bayesian implementation supports random effects to control for one-layer clustering. In this paper, we give an overview of the methodology and illustrate the package's functionalities using a veterinary dataset about respiratory diseases in commercial swine production.

stat.ML↗

arXiv2019

Combining Heterogeneous Spatial Datasets with Process-based Spatial Fusion Models: A Unifying Framework

In modern spatial statistics, the structure of data that is collected has become more heterogeneous. Depending on the type of spatial data, different modeling strategies for spatial data are used. For example, a kriging approach for geostatistical data; a Gaussian Markov random field model for lattice data; or a log Gaussian Cox process for point-pattern data. Despite these different modeling choices, the nature of underlying scientific data-generating (latent) processes is often the same, which can be represented by some continuous spatial surfaces. In this paper, we introduce a unifying framework for process-based multivariate spatial fusion models. The framework can jointly analyze all three aforementioned types of spatial data (or any combinations thereof). Moreover, the framework accommodates different conditional distributions for geostatistical and lattice data. We show that some established approaches, such as linear models of coregionalization, can be viewed as special cases of our proposed framework. We offer flexible and scalable implementations in R using Stan and INLA. Simulation studies confirm that the predictive performance of latent processes improves as we move from univariate spatial models to multivariate spatial fusion models. The introduced framework is illustrated using a cross-sectional study linked with a national cohort dataset in Switzerland, we examine differences in underlying spatial risk patterns between respiratory disease and lung cancer.

stat.ME↗

arXiv2019

Is a single unique Bayesian network enough to accurately represent your data?

Bayesian network (BN) modelling is extensively used in systems epidemiology. Usually it consists in selecting and reporting the best-fitting structure conditional to the data. A major practical concern is avoiding overfitting, on account of its extreme flexibility and its modelling richness. Many approaches have been proposed to control for overfitting. Unfortunately, they essentially all rely on very crude decisions that result in too simplistic approaches for such complex systems. In practice, with limited data sampled from complex system, this approach seems too simplistic. An alternative would be to use the Monte Carlo Markov chain model choice (MC3) over the network to learn the landscape of reasonably supported networks, and then to present all possible arcs with their MCMC support. This paper presents an R implementation, called mcmcabn, of a flexible structural MC3 that is accessible to non-specialists.

stat.CO↗

arXiv2018

Comparison between Suitable Priors for Additive Bayesian Networks

Additive Bayesian networks are types of graphical models that extend the usual Bayesian generalized linear model to multiple dependent variables through the factorisation of the joint probability distribution of the underlying variables. When fitting an ABN model, the choice of the prior of the parameters is of crucial importance. If an inadequate prior - like a too weakly informative one - is used, data separation and data sparsity lead to issues in the model selection process. In this work a simulation study between two weakly and a strongly informative priors is presented. As weakly informative prior we use a zero mean Gaussian prior with a large variance, currently implemented in the R-package abn. The second prior belongs to the Student's t-distribution, specifically designed for logistic regressions and, finally, the strongly informative prior is again Gaussian with mean equal to true parameter value and a small variance. We compare the impact of these priors on the accuracy of the learned additive Bayesian network in function of different parameters. We create a simulation study to illustrate Lindley's paradox based on the prior choice. We then conclude by highlighting the good performance of the informative Student's t-prior and the limited impact of the Lindley's paradox. Finally, suggestions for further developments are provided.

stat.ME↗

arXiv2018

Information-Theoretic Scoring Rules to Learn Additive Bayesian Network Applied to Epidemiology

Bayesian network modelling is a well adapted approach to study messy and highly correlated datasets which are very common in, e.g., systems epidemiology. A popular approach to learn a Bayesian network from an observational datasets is to identify the maximum a posteriori network in a search-and-score approach. Many scores have been proposed both Bayesian or frequentist based. In an applied perspective, a suitable approach would allow multiple distributions for the data and is robust enough to run autonomously. A promising framework to compute scores are generalized linear models. Indeed, there exists fast algorithms for estimation and many tailored solutions to common epidemiological issues. The purpose of this paper is to present an R package abn that has an implementation of multiple frequentist scores and some realistic simulations that show its usability and performance. It includes features to deal efficiently with data separation and adjustment which are very common in systems epidemiology.

stat.ML↗

arXiv2018

optimParallel: an R Package Providing Parallel Versions of the Gradient-Based Optimization Methods of optim()

The R package optimParallel provides a parallel version of the gradient-based optimization methods of optim(). The main function of the package is optimParallel(), which has the same usage and output as optim(). Using optimParallel() can significantly reduce optimization times. We introduce the R package and illustrate its implementation, which takes advantage of the lexical scoping mechanism of R.

stat.CO↗

arXiv2018

A Case Study Competition Among Methods for Analyzing Large Spatial Data

The Gaussian process is an indispensable tool for spatial data analysts. The onset of the "big data" era, however, has lead to the traditional Gaussian process being computationally infeasible for modern spatial data. As such, various alternatives to the full Gaussian process that are more amenable to handling big spatial data have been proposed. These modern methods often exploit low rank structures and/or multi-core and multi-threaded computing environments to facilitate computation. This study provides, first, an introductory overview of several methods for analyzing large spatial data. Second, this study describes the results of a predictive competition among the described methods as implemented by different groups with strong expertise in the methodology. Specifically, each research group was provided with two training datasets (one simulated and one observed) along with a set of prediction locations. Each group then wrote their own implementation of their method to produce predictions at the given location and each which was subsequently run on a common computing environment. The methods were then compared in terms of various predictive diagnostics. Supplementary materials regarding implementation details of the methods and code are available for this article online.

stat.ME↗

arXiv2018

varrank: an R package for variable ranking based on mutual information with applications to observed systemic datasets

This article describes the R package varrank. It has a flexible implementation of heuristic approaches which perform variable ranking based on mutual information. The package is particularly suitable for exploring multivariate datasets requiring a holistic analysis. The core functionality is a general implementation of the minimum redundancy maximum relevance (mRMRe) model. This approach is based on information theory metrics. It is compatible with discrete and continuous data which are discretised using a large choice of possible rules. The two main problems that can be addressed by this package are the selection of the most representative variables for modeling a collection of variables of interest, i.e., dimension reduction, and variable ranking with respect to a set of variables of interest.

stat.ML↗

arXiv2017

Covariance Functions for Multivariate Gaussian Fields Evolving Temporally over Planet Earth

The construction of valid and flexible cross-covariance functions is a fundamental task for modeling multivariate space-time data arising from climatological and oceanographical phenomena. Indeed, a suitable specification of the covariance structure allows to capture both the space-time dependencies between the observations and the development of accurate predictions. For data observed over large portions of planet Earth it is necessary to take into account the curvature of the planet. Hence the need for random field models defined over spheres across time. In particular, the associated covariance function should depend on the geodesic distance, which is the most natural metric over the spherical surface. In this work, we propose a flexible parametric family of matrix-valued covariance functions, with both marginal and cross structure being of the Gneiting type. We additionally introduce a different multivariate Gneiting model based on the adaptation of the latent dimension approach to the spherical context. Finally, we assess the performance of our models through the study of a bivariate space-time data set of surface air temperatures and precipitations.

math.ST↗

arXiv2017

Modeling Temporally Evolving and Spatially Globally Dependent Data

The last decades have seen an unprecedented increase in the availability of data sets that are inherently global and temporally evolving, from remotely sensed networks to climate model ensembles. This paper provides a view of statistical modeling techniques for space-time processes, where space is the sphere representing our planet. In particular, we make a distintion between (a) second order-based, and (b) practical approaches to model temporally evolving global processes. The former are based on the specification of a class of space-time covariance functions, with space being the two-dimensional sphere. The latter are based on explicit description of the dynamics of the space-time process, i.e., by specifying its evolution as a function of its past history with added spatially dependent noise. We especially focus on approach (a), where the literature has been sparse. We provide new models of space-time covariance functions for random fields defined on spheres cross time. Practical approaches, (b), are also discussed, with special emphasis on models built directly on the sphere, without projecting the spherical coordinate on the plane. We present a case study focused on the analysis of air pollution from the 2015 wildfires in Equatorial Asia, an event which was classified as the year's worst environmental disaster. The paper finishes with a list of the main theoretical and applied research problems in the area, where we expect the statistical community to engage over the next decade.

math.ST↗

arXiv2017

Asymmetric Matrix-Valued Covariances for Multivariate Random Fields on Spheres

Matrix-valued covariance functions are crucial to geostatistical modeling of multivariate spatial data. The classical assumption of symmetry of a multivariate covariance function is overlay restrictive and has been considered as unrealistic for most of real data applications. Despite of that, the literature on asymmetric covariance functions has been very sparse. In particular, there is some work related to asymmetric covariances on Euclidean spaces, depending on the Euclidean distance. However, for data collected over large portions of planet Earth, the most natural spatial domain is a sphere, with the corresponding geodesic distance being the natural metric. In this work, we propose a strategy based on spatial rotations to generate asymmetric covariances for multivariate random fields on the $d$-dimensional unit sphere. We illustrate through simulations as well as real data analysis that our proposal allows to achieve improvements in the predictive performance in comparison to the symmetric counterpart.

math.ST↗

arXiv2017

dotCall64: An Efficient Interface to Compiled C/C++ and Fortran Code Supporting Long Vectors

The R functions .C() and .Fortran() can be used to call compiled C/C++ and Fortran code from R. This so-called foreign function interface is convenient, since it does not require any interactions with the C API of R. However, it does not support long vectors (i.e., vectors of more than 2^31 elements). To overcome this limitation, the R package dotCall64 provides .C64(), which can be used to call compiled C/C++ and Fortran functions. It transparently supports long vectors and does the necessary castings to pass numeric R vectors to 64-bit integer arguments of the compiled code. Moreover, .C64() features a mechanism to avoid unnecessary copies of function arguments, making it efficient in terms of speed and memory usage.

stat.CO↗

arXiv2016

Predicting missing values in spatio-temporal satellite data

Remotely sensed data are sparse, which means that data have missing values, for instance due to cloud cover. This is problematic for applications and signal processing algorithms that require complete data sets. To address the sparse data issue, we present a new gap-fill algorithm. The proposed method predicts each missing value separately based on data points in a spatio-temporal neighborhood around the missing data point. The computational workload can be distributed among several computers, making the method suitable for large datasets. The prediction of the missing values and the estimation of the corresponding prediction uncertainties are based on sorting procedures and quantile regression. The algorithm was applied to MODIS NDVI data from Alaska and tested with realistic cloud cover scenarios featuring up to 50% missing data. Validation against established software showed that the proposed method has a good performance in terms of the root mean squared prediction error. The procedure is implemented and available in the open-source R package gapfill. We demonstrate the software performance with a real data example and show how it can be tailored to specific data. Due to the flexible software design, users can control and redesign major parts of the procedure with little effort. This makes it an interesting tool for gap-filling satellite data and for the future development of gap-fill procedures.

stat.AP↗

arXiv2016

Valid parameter space of a bivariate Gaussian Markov random field with a generalized block-Toeplitz precision matrix

Gaussian Markov random fields (GMRFs) are extensively used in statistics to model area-based data and usually depend on several parameters in order to capture complex spatial correlations. In this context, it is important to determine the valid parameter space, namely the domain ensuring (semi) positive-definiteness of the precision matrix. Depending on the structure of the latter, this task can be challenging. While univari- ate GMRFs with block-Toeplitz precision are well studied in the literature, not much is analytically known about bivariate GMRFs. So far, only restrictive sufficient conditions and brute-force approaches were proposed, which are computationally expensive for the size of modern datasets. In this paper, we consider a bivariate GMRF, which is part of a hierarchical model used in spatial statistics to analyze data coming from projec- tions of regional climate change. By extending classical convergence results of univariate fields with toroidal boundary conditions to fields without boundary conditions, we pro- vide asymptotically closed-form expressions of the valid parameter space. We develop a general methodology that can be used to determine the valid parameter space of bivariate GMRFs whose precision matrix has a generalized block-Toeplitz structure and for which classical convergence results are not directly applicable. Finally, we quantify the rate of convergence of our approach through a numerical study in R.

stat.CO↗