University of Newcastle upon Tyne

School of Mathematics and Statistics

Statistics Seminars 2005-2006

4 November 2005, L401, 2:15pm

Dr Ruth King, University of St Andrews

Abstract

The use of Bayesian methods within the field of Ecology is becoming increasingly widespread. We consider a particular application to capture-recapture data and use a real example relating to a population of Soay sheep, for illustration. Within the data collection process, additional covariate information is often collected (possibly with some missing values), which can be incorporated into the analysis. Then, typically, not only is parameter estimation of interest, but also the identification of the underlying model, i.e. in our case the covariates that influence the survival rates of the population under study. This can be very important in understanding the underlying dynamics of the population as well as making predictive inference. We consider a Bayesian approach, where we are able to quantitatively discriminate between competing models by their corresponding posterior model probabilities, and hence obtain posterior model-averaged estimates of interest. These are obtained using the reversible jump MCMC algorithm, and we discuss the interpretation of the results that are obtained.

Return to timetable for 2005-06 seminars

School of Mathematics and Statistics Homepage