ARCHIVED | SASNet Seminar | Mortality Estimation and Forecasting With Smoothing and Overdispersion

A review of ‘Empowering the Public Sector: Delving into data analytics’
2 February, 2017
ARCHIVED | Event | Seminar Series | Statistics and Data Science in the Digital Age
16 February, 2017
Show all
0

ARCHIVED | SASNet Seminar | Mortality Estimation and Forecasting With Smoothing and Overdispersion

Social Analytics Strategic Network Seminar

Archived – Please note this event took place in the past and has been left here for reference purposes. View our full list of events to see what we have coming up or if there is a particular type of event you are interested in.

Mortality Estimation and Forecasting With Smoothing and Overdispersion

Date: 14 March
Time: 2pm to 3:30pm (with networking afterwards)
Venue: Colchester Campus, University of Essex

The ESRC Business and Local Government Data Research Centre in conjunction with SASNet invites you to participate in a free training seminar on Mortality Estimation and Forecasting With Smoothing and Overdispersion, featuring Peter W. F. Smith, Professor of Social Statistics at the University of Southampton.

Abstract
Peter W. F. Smith, Professor of Social Statistics at the University of Southampton, will deliver a seminar detailing proposals of a comprehensive mortality modelling framework, which overcomes several of the limitations associated with existing approaches.

In this training course, Peter W. F. Smith will discuss the current issues with existing models as outlined below, and explain in greater detail how the proposed modelling framework accounts for these. This seminar is aimed at attendees who are interested in estimating and forecasting mortality rates as well as statistical modelling or population processes more generally. Some prior knowledge of statistical modelling would be useful, as Peter will present some model equations, although he will explain the ideas in graphs and words as well.

Background
Often, error models do not adequately account for the variability in the data which can lead to estimates and forecasts being over-fit and insufficiently robust. Peter will explain how the proposed mortality modelling framework accounts for this by specifying a negative binomial error structure.

Another feature, lacking in many existing approaches, is the facility to impose smoothness in parameter series which vary over age, cohort and time. Such constraints are integrated into the modelling process, so that there is a natural feedback, whereby the smoothing of parameter series can appropriately impact other estimates, rather than being performed in a post hoc fashion. Peter will discuss how the proposed modelling framework demonstrates that generalised additive models (GAMs) can be used for mortality modelling and forecasting. GAMs are a flexible class of semiparametric statistical models which allow parametric functions and unstructured (but smooth) functions of explanatory variables to appear in the model simultaneously. In particular, GAMs allow us to differentially smooth components, such as cohorts, more aggressively in areas of sparse data for the component concerned. While GAMs can provide a reasonable fit for the ages where there is adequate data, estimation and extrapolation of mortality rates using a GAM at higher ages is problematic due to high variation in crude rates. At these ages, parametric models can give a more robust fit, enabling a borrowing of strength across age groups. Peter and colleague’s modelling methodology is based on a smooth transition between a GAM at lower ages and a fully parametric model at higher ages. Finally, their framework is fully probabilistic, and provides a coherent description of forecast uncertainty.

Presenter Biography
Peter W. F. Smith is Professor of Social Statistics at the University of Southampton. He also is Director of the Administrative Data Research Centre for England. Peter has worked at the University for over 25 years. He obtained a First Class BSc in Mathematics in 1986 from Lancaster University, and returned there to complete a PhD in Statistics in 1990, having obtained an MSc in Probability and Statistics with Distinction in 1987 from the University of Sheffield. Peter has research interests in developing new statistical methodology, including methods for handling non-response and for modelling mortality data, and applying sophisticated statistical methods to problems in demography, medicine and health sciences. His publications include articles in the Journal of the Royal Statistical Society, Series A, B and C, Biometrika and the Journal of the American Statistical Association. Peter was awarded the Royal Statistical Society Guy Medal in Bronze in 1999 and was Joint Editor of Series C of their Journal from 2013 to 2016.

Book now

***Please note: Registration is required for catering purposes. Refreshments will be provided.