Discussion meetings

Discussion Meetings are events where articles ('papers for reading') appearing in the Journal of the Royal Statistical Society are presented and discussed. The discussion and authors' replies are then published in the relevant Journal series. About half of the meetings are organised by the Society's Research Section and the events are often preceded by an informal session on the issues raised by the papers. See our guidelines for papers for discussion.

Discussion Meetings are usually held in the early evening at the Society's premises in Errol Street, London. Any change in venue will be noted alongside the meeting details.

Preprints of journal papers are available to download to encourage discussion at our Discussion Meetings before publication in one of our journals. Other papers, such as Presidential addresses, are also available to download. All preprints available here are provisional and subject to later amendment by the authors.

Contact Judith Shorten if you would like to make a written contribution to a discussion meeting or receive a preprint for each meeting by email.

Click here to watch videos from past discussion meetings.

Preprint discussion papers

2019

Research Section Discussion Meeting, Wednesday, 16 October 2019
‘Functional models for time-varying random objects’
Paromita Dubey and Hans-Georg Müller
Details

Discussion Meeting, Wednesday, 13 November 2019
‘Multiple-systems analysis for the quantification of modern slavery: classical and Bayesian approaches’
Bernard W. Silverman
Details

Research Section Discussion Meeting, Wednesday, 11 December 2019
‘Unbiased Markov chain Monte Carlo methods with couplings’
Pierre E. Jacob, John O’Leary and Yves F. Atchadé
Details


Preprints

2019

Research Section Discussion Meeting, Wednesday, 16 October 2019
Paromita Dubey and Hans-Georg Müller (University of California at Davis, USA)
‘Functional models for time-varying random objects

Functional data analysis provides a popular toolbox of functional models for the analysis of samples of random functions that are real valued. In recent years, samples of time-varying object data such as time-varying networks that are not in a vector space have been increasingly collected. These data can be viewed as elements of a general metric space that lacks local or global linear structure and therefore common approaches that have been used with great success for the analysis of functional data, such as functional principal component analysis, cannot be applied. We propose metric covariance, a novel association measure for paired object data lying in a metric space (Ω, d) that we use to define a metric autocovariance function for a sample of random Ω-valued curves, where Ω generally will not have a vector space or manifold structure. The proposed metric autocovariance function is non-negative definite when the squared semimetric d2 is of negative type. Then the eigenfunctions of the linear operator with the autocovariance function as kernel can be used as building blocks for an object functional principal component analysis for Ω-valued functional data, including time-varying probability distributions, covariance matrices and time dynamic networks. Analogues of functional principal components for time-varying objects are obtained by applying Fréchet means and projections of distance functions of the random object trajectories in the directions of the eigenfunctions, leading to real-valued Fréchet scores. Using the notion of generalized Fréchet integrals, we construct object functional principal components that lie in the metric space Ω. We establish asymptotic consistency of the sample-based estimators for the corresponding population targets under mild metric entropy conditions on Ω and continuity of the Ω-valued random curves. These concepts are illustrated with samples of time-varying probability distributions for human mortality, time-varying covariance matrices derived from trading patterns and time-varying networks that arise from New York taxi trips.

To be published in Series B; for more information go to the Wiley Online Library.
The preprint is available to download. ‘Functional models for time-varying random objects’ (PDF)
Supporting information (PDF)
Movies


Discussion Meeting, Wednesday, 13 November 2019
Bernard W. Silverman (University of Nottingham, UK)
‘Multiple-systems analysis for the quantification of modern slavery: classical and Bayesian approaches’

Multiple-systems estimation is a key approach for quantifying hidden populations such as the number of victims of modern slavery. The UK Government published an estimate of 10000–13000 victims, constructed by the present author, as part of the strategy leading to the Modern Slavery Act 2015. This estimate was obtained by a stepwise multiple-systems method based on six lists. Further investigation shows that a small proportion of the possible models give rather different answers, and that other model fitting approaches may choose one of these. Three data sets collected in the field of modern slavery, together with a data set about the death toll in the Kosovo conflict, are used to investigate the stability and robustness of various multiple-systems-estimate methods. The crucial aspect is the way that interactions between lists are modelled, because these can substantially affect the results. Model selection and Bayesian approaches are considered in detail, in particular to assess their stability and robustness when applied to real modern slavery data. A new Markov chain Monte Carlo Bayesian approach is developed; overall, this gives robust and stable results at least for the examples considered. The software and data sets are freely and publicly available to facilitate wider implementation and further research.

To be published in Series A; for more information go to the Wiley Online Library.

The preprint is available to download.
Multiple-systems analysis for the quantification of modern slavery: classical and Bayesian approaches’ (PDF)
Data set and computer code (ZIP)


Research Section Discussion Meeting, Wednesday, 11 December 2019
Pierre E. Jacob and John O’Leary (Harvard University, Cambridge) and Yves F. Atchadé (Boston University)
‘Unbiased Markov chain Monte Carlo methods with couplings’

Markov chain Monte Carlo (MCMC) methods provide consistent approximations of integrals as the number of iterations goes to 1. MCMC estimators are generally biased after any fixed number of iterations. We propose to remove this bias by using couplings of Markov chains together with a telescopic sum argument of Glynn and Rhee. The resulting unbiased estimators can be computed independently in parallel. We discuss practical couplings for popular MCMC algorithms. We establish the theoretical validity of the estimators proposed and study their efficiency relative to the underlying MCMC algorithms. Finally, we illustrate the performance and limitations of the method on toy examples, on an Ising model around its critical temperature, on a high dimensional variable-selection problem, and on an approximation of the cut distribution arising in Bayesian inference for models made of multiple modules.

To be published in Series B; for more information go to the Wiley Online Library.

The preprint is available to download.
Unbiased Markov chain Monte Carlo methods with couplings’ (PDF)
Supporting information (PDF)