Use this URL to cite or link to this record in EThOS:
Title: Bayesian model discrimination with application to population ecology and epidemiology
Author: Jamieson, L. E.
Awarding Body: University of Cambridge
Current Institution: University of Cambridge
Date of Award: 2004
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
The development of Bayesian methods has flourished in recent years, due in part to the surge in computer power and the development of algorithmic tools. In particular Markov chain Monte Carlo (MCMC) has revolutionised this field and the exciting recent advance of reversible jump MCMC (RJMCMC) allows for greater across model inference. This multimodel sampler provides a cohesive framework for model selection and model averaging. In this thesis we demonstrate how these powerful new techniques can be used to discriminate between competing biological hypotheses in two application areas, population ecology and epidemiology. National and International authorities are becoming increasingly concerned with the management and maintenance of key wildlife species and their natural habitats. Central to any management programme is the construction of realistic models of the underlying population dynamics and their interaction with the local environment. We focus upon the analysis of a series of data on North American ducks and construct a state space model to investigate the importance of population size and other key covariates on the underlying population dynamics. In epidemiology understanding the dynamics of how an infectious disease spreads is vital for both eradication and containment. In modelling a disease outbreak we are interested in both the spatial scale (how the spread of infection is affected by distance) and the temporal scale (how quickly an individual is infected). The spatial scale provides invaluable information on methods to contain the epidemic, while the temporal scale will tell us when the containment measures should be assessed for efficacy. We apply the methodology to two plant diseases: a vector-borne viral disease Citrus Tristeza and a bacterial pathogen Citrus Canker.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available