Bayesian inference from composite likelihoods, with an application to spatial extremes

Anthony Christopher Davison
Academia Sinica, Institute of Statistical Science, 2012
Article

Résumé

Composite likelihoods are increasingly used in applications where the full likelihood is analytically unknown or computationally prohibitive. Although some frequentist properties of the maximum composite likelihood estimator are akin to those of the maximum likelihood estimator, Bayesian inference based on composite likelihoods is in its early stages. This paper discusses inference when one uses composite likelihood in Bayes' formula. We establish that using a composite likelihood results in a proper posterior density, though it can differ considerably from that stemming from the full likelihood. Building on previous work on composite likelihood ratio tests, we use asymptotic theory for misspecified models to propose two adjustments to the composite likelihood to obtain appropriate inference. We also investigate use of the Metropolis Hastings algorithm and two implementations of the Gibbs sampler for obtaining draws from the composite posterior. We test the methods on simulated data and apply them to a spatial extreme rainfall dataset. For the simulated data, we find that posterior credible intervals yield appropriate empirical coverage rates. For the extreme precipitation data, we are able to both effectively model marginal behavior throughout the study region and obtain appropriate measures of spatial dependence.

Official source

https://infoscience.epfl.ch/record/173427?ln=fr

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Concepts associés

Chargement

Publications associées

Chargement

xtreme value analysis is concerned with the modelling of extreme events such as floods and heatwaves, which can have large impacts. Statistical modelling can be useful to better assess risks even if, due to scarcity of measurements, there is inherently very large residual uncertainty in any analysis. Driven by the increase in environmental databases, spatial modelling of extremes has expanded rapidly in the last decade. This thesis presents contributions to such analysis. The first chapter is about likelihood-based inference in the univariate setting and investigates the use of bias-correction and higher-order asymptotic methods for extremes, highlighting through examples and illustrations the unique challenge posed by data scarcity. We focus on parametric modelling of extreme values, which relies on limiting distributional results and for which, as a result, uncertainty quantification is complicated. We find that, in certain cases, small-sample asymptotic methods can give improved inference by reducing the error rate of confidence intervals. Two data illustrations, linked to assessment of the frequency of extreme rainfall episodes in Venezuela and the analysis of survival of supercentenarians, illustrate the methods developed. In the second chapter, we review the major methods for the analysis of spatial extremes models. We highlight the similarities and provide a thorough literature review along with novel simulation algorithms. The methods described therein are made available through a statistical software package. The last chapter focuses on estimation for a Bayesian hierarchical model derived from a multivariate generalized Pareto process. We review approaches for the estimation of censored components in models derived from (log)-elliptical distributions, paying particular attention to the estimation of a high-dimensional Gaussian distribution function via Monte Carlo methods. The impacts of model misspecification and of censoring are explored through extensive simulations and we conclude with a case study of rainfall extremes in Eastern Switzerland.

EPFL2019

Chargement

Extreme events are responsible for huge material damage and are costly in terms of their human and economic impacts. They strike all facets of modern society, such as physical infrastructure and insurance companies through environmental hazards, banking and finance through stock market crises, and the internet and communication systems through network and server overloads. It is thus of increasing importance to accurately assess the risk of extreme events in order to mitigate them. Extreme value theory is a statistical approach to extrapolation of probabilities beyond the range of the data, which provides a robust framework to learn from an often small number of recorded extreme events. In this thesis, we consider a conditional approach to modelling extreme values that is more flexible than standard models for simultaneously extreme events. We explore the subasymptotic properties of this conditional approach and prove that in specific situations its finite-sample behaviour can differ significantly from its limit characterisation. For modelling extremes in time series with short-range dependence, the standard peaks-over-threshold method relies on a pre-processing step that retains only a subset of observations exceeding a high threshold and can result in badly-biased estimates. This method focuses on the marginal distribution of the extremes and does not estimate temporal extremal dependence. We propose a new methodology to model time series extremes using Bayesian semiparametrics and allowing estimation of functionals of clusters of extremes. We apply our methodology to model river flow data in England and improve flood risk assessment by explicitly describing extremal dependence in time, using information from all exceedances of a high threshold. We develop two new bivariate models which are based on the conditional tail approach, and use all observations having at least one extreme component in our inference procedure, thus extracting more information from the data than existing approaches. We compare the efficiency of these models in a simulation study and discuss generalisations to higher-dimensional setups. Existing models for extremes of Markov chains generally rely on a strong assumption of asymptotic dependence at all lags and separately consider marginal and joint features. We introduce a more flexible model and show how Bayesian semiparametrics can provide a suitable framework allowing simultaneous inference for the margins and the extremal dependence structure, yielding efficient risk estimates and a reliable assessment of uncertainty.

EPFL2018

Chargement

This thesis focuses on two kinds of statistical inference problems in signal processing and data science. The first problem is the estimation of a structured informative tensor from the observation of a noisy tensor in which it is buried. The structure comes from the possibility to decompose the informative tensor as the sum of a small number of rank-one tensors (small compared to its size). Such structure has applications in data science where data, organized into arrays, can often be explained by the interaction between a few features characteristic of the problem. The second problem is the estimation of a signal input to a feedforward neural network whose output is observed. It is relevant for many applications (phase retrieval, quantized signals) where the relation between the measurements and the quantities of interest is not linear. We look at these two statistical models in different high-dimensional limits corresponding to situations where the amount of observations and size of the signal become infinitely large. These asymptotic regimes are motivated by the ever-increasing computational power and storage capacity that make possible the processing and handling of large data sets. We take an information-theoretic approach in order to establish fundamental statistical limits of estimation in high-dimensional regimes. In particular, the main contributions of this thesis are the proofs of exact formulas for the asymptotic normalized mutual information associated with these inference problems. These are low-dimensional variational formulas that can nonetheless capture the behavior of a large system where each component interacts with all the others. Owing to the relationship between mutual information and Bayesian inference, we use the solutions to these variational problems to rigorously predict the asymptotic minimum mean-square error (MMSE), the error achieved by the (Bayes) optimal estimator. We can thus compare algorithmic performances to the statistical limit given by the MMSE. Variational formulas for the mutual information are referred to as replica symmetric (RS) ansätze due to the predictions of the heuristic replica method from statistical physics. In the past decade proofs of the validity of these predictions started to emerge. The general strategy is to show that the RS ansatz is both an upper and lower bound on the asymptotic normalized mutual information. The present work leverages on the adaptive interpolation method that proposes a unified way to prove the two bounds. We extend the adaptive interpolation to situations where the order parameter of the problem is not a scalar but a matrix, and to high-dimensional regimes that differ from the one for which the RS formula is usually conjectured. Our proofs also demonstrate the modularity of the method. Indeed, using statistical models previously studied in the literature as building blocks of more complex ones (e.g., estimated signal with a model of structure), we derive RS formulas for the normalized mutual information associated with estimation problems that are relevant to modern applications.

EPFL2021

Chargement

EPFL2019

EPFL2021

EPFL2018