SEMINAR: The University of Reading,
School of Applied Statistics, room AS-G03
Thursday, June 24, 2pm
William Browne (University of Nottingham)
An MCMC algorithm for problems involving 'constrained' variance
matrices with applications in multilevel modelling
General-purpose Bayesian software packages, for example WinBUGS
(Spiegelhalter et al. 2000) that utilize MCMC methods are now
being used widely by quantitative researchers. To make such
software as flexible as possible, MCMC methods that can be
adapted to fit the widest range of statistical models have
been preferred. Originally, Gibbs sampling algorithms primarily
through the AR sampler (Gilks and Wild 1992) were used to fit
models using univariate updates with the restriction that all
conditional posteriors be log concave. More recently this
restriction has been removed by using adaptive (random-walk)
Metropolis samplers for parameters without log concave distributions.
These samplers are used (where necessary) in both WinBUGS and MLwiN
software packages (Rasbash et al. 2000). In this talk we discuss
another feature of certain statistical models, 'constrained'
variance matrices, which cannot currently be dealt with in
general purpose packages. By 'constrained' we mean that the
variance matrix is subject to some additional constraints
(as well as the positive definite constraint). For example,
two elements of the matrix could be constrained to be equal,
or an element could equal a constant or be a function of
predictor variables.
Full abstract on web-page below
All welcome.
Travel information: http://www.rdg.ac.uk/Statistics/info/direc.html
Other seminars: http://www.rdg.ac.uk/Statistics/diary/seminars.html
Posted by: Howard Grubb ([log in to unmask])
|