Tag Archives: Conferences

New papers submitted to Chaire Quetelet 2013

I’m excited to be involved in the upcoming Chaire Quetelet 2013, a seminar focusing on changes in the field of demography over the last 50 years and how we see the field changing in the next 50.  Many well-known faces from the field will be present for the discussion — some speaking in English, others speaking in French! — and it will certainly be a stimulating and challenging forum.

I’ve had the pleasure of collaborating on two papers for the seminar with Daniel Courgeau, Robert Franck, and Jakub Bijak.  We’ve recently submitted the final papers and you can access them in PDF form here:

Quantifying paradigm change in demography

Are the four Baconian Idols still alive in demography?

Feel free to download, ponder and send your comments!  They’re available on my ResearchGate and Academia.edu pages as well.

Tagged , , , , ,

Presenting at Eurostat / UNECE Work Session on Population Projections

We got some pleasing news on Friday: our abstract entitled From agent-based models to statistical emulators has been accepted for presentation at the Eurostat/UNECE Work Session on Population Projections in Rome, Italy from 29 – 31 October 2013.   This will be a great opportunity for me to link up with more demographers and gain greater exposure to that community.  As ever I’m curious to find out how our unconventional methods of modelling will be received!

Our abstract is below:

From agent-based models to statistical emulators

Jakub Bijak, Jason Hilton and Eric Silverman

University of Southampton, Southampton, SO17 1BJ, United Kingdom

Contact author: j.bijak@soton.ac.uk

Proposed for the strand on «New methodologies» Eurostat / UNECE Work Session on Population Projections; Rome, 29–31 October 2013

Contemporary demographic micro-simulations are largely concerned with populations of statistical individuals, whose life courses can be inferred from empirical information (Courgeau 2012). In contrast, agent-based models study simulated individuals, for whom certain behavioural rules are assumed. We wish to bring these two approaches closer together by coupling the rule-based explanations driving the agent-based model with observed data. We also propose a method to analyse the statistical properties of such models, based on the notion of statistical emulators (Kennedy & O’Hagan 2001; Oakley & O’Hagan 2002).

In this paper, we present a Semi-Artificial Model of Population, which aims to bridge demographic micro-simulation and agent-based traditions. We extend the ‘Wedding Ring’ agent-based model of marriage formation (Billari et al. 2007) to include empirical information on the natural population change for the United Kingdom alongside with the behavioural explanations that drive the observed demographic trends. The mortality and fertility rates in this population are drawn from UK population data for 1951–2011 and forecasts until 2250 obtained from Lee-Carter models. We then utilise a Gaussian process emulator – a statistical model of the base model – to analyse the impact of selected parameters on two key simulation outputs: population size and share of agents with partners. A sensitivity analysis is attempted, aiming to assess the relative importance of different inputs.

The resulting multi-state model of population dynamics is argued to have enhanced predictive capacity as compared to the original specification of the Wedding Ring, but there are some trade-offs between the outputs considered. The sensitivity analysis indicates a key role of social pressure in the modelled partnership formation process. We posit that the presented method allows for generating coherent, multi-level agent-based scenarios aligned with selected aspects of empirical demographic reality. Emulators permit a statistical analysis of the model properties and help select plausible parameter values. Given non-linearities in agent-based models such as the Wedding Ring, and the presence of feedback loops, the uncertainty of the model may be impossible to assess directly with traditional statistical methods. The use of statistical emulators offers a way forward.

Billari, F., Aparicio Diaz, B., Fent, T. and Prskawetz, A. (2007) The “Wedding–Ring”. An agent–based marriage model based on social interaction. Demographic Research, 17(3): 59–82.

Courgeau, D. (2012). Probability and Social Science. Methodological Relationships between the two Approaches. Dordrecht: Springer.

Kennedy, M., and O’Hagan, T. (2001) Bayesian Calibration of Computer Models. Journal of the Royal Statistical Society, Series B, 63(3), pp. 425–464.

Oakley, J. and O’Hagan, A. (2002) Bayesian inference for the uncertainty distribution of computer model outputs. Biometrika, 89(4), pp. 769–784.

Tagged ,

Paper accepted to ECMS 2013

I’ve just had a paper accepted to the European Conference on Modelling and Simulation 2013 which is being held from 27-30 May in Ålesund, Norway.  The paper is titled Simulating the Cost of Social Care in an Ageing Population and was written by myself, Jason Hilton, Jason Noble, and Jakub Bijak.  We were accepted to the Policy Modelling track, so I’m hoping for some interesting feedback from other researchers who may be working on projects aimed at health and social care.

The reviews were very positive on the whole, so we’re pleased about that!  Corrections are still to come before the paper enters the Proceedings, but in the meantime you can find the submitted draft here.

Tagged , , ,

Submitted a paper to IUSSP

Just yesterday my colleagues Jakub Bijak and Jason Hilton and myself submitted a paper (4-page extended abstract, with a paper to follow if accepted) to the XXVII IUSSP Population Conference, to be held in Busan, South Korea next August.  The paper is titled “Statistical Individuals and Simulated Individuals: Analysing Agent-Based Demographic Models with Gaussian Process Emulators“, which is a very long title… need to take a rest after saying all that.  The content is essentially a distillation of our recent work on agent-based models for the study of population dynamics, with an emphasis on the potential impact of this type of simulation methodology on event-history analysis.

The conference is run by the IUSSP, the International Union for the Scientific Study of Population, based in Paris.  The attendees will be 2-3,000 demographers across all areas of the discipline, so this will hopefully be my first opportunity to be on hand to present our recent work to the demography community.  I’m very interested to see how these conferences work; my field is quite small in comparison, and our big conferences are perhaps 1/10th the size of this one!

Tagged , , ,

Back in Southampton

I’ve returned to the UK now after a lengthy sojourn in East Asia, which turned out to be genuinely life-changing.  Following an exciting and interesting conference in Taipei, I traveled to Tokyo, where I got married (!), and enjoyed 3 fantastic weeks of relaxation and exploration with my partner.

Future posts will get back into more academic matters as I settle into my normal patterns once again.  I suspect there may be some exciting research developments on hand in the near future….

Tagged ,