Please see below for details of the following WinBUGS and other Bayesian
courses that we are running at Imperial College later this year.
WinBUGS Courses
Course 1: (Run in conjunction with the RSS Environmental Statistics Study
Group)
Bayesian models in environmental applications using WinBUGS
1.5 day course Thursday June 12 and Friday June 13, 2003
Course 2:
Bayesian Hierarchical Models
3 day course Wed September 3 - Friday September 5 2003
Course 3:
Spatial Epidemiology
2 day course Mon September 8 - Tues September 9 2003
Other courses
Introduction to genetic epidemiology
2-day course Wed Sept 10 - Thurs Sept 11 2003
Statistical analysis of genetic association studies
1-day course Fri Sept 12 2003
Nicky Best
------------------------------------------
ESSG workshop on
Bayesian models in environmental applications using WinBUGS
Thursday June 12 and 13, 2003, Imperial College, London
Begins 9am Thursday, ends 3pm Friday
Course presenters: Dr Nicky Best and Prof Sylvia Richardson
Course overview: Bayesian methods are becoming increasingly popular in a
remarkably wide range of applied data analysis problems. One reason for this
is the flexibility of modelling made possible by the simulation-based
methods, such as Markov chain Monte Carlo (MCMC), used for Bayesian
estimation. Environmental and spatial/temporal data are particularly well
suited to this approach since MCMC methods allow realistic models to be
fitted. Such models can readily incorporate features such as spatial and/or
temporal structure, different geographical scales of measurement, missing
data, measurement error and so on. This course will introduce participants
to a range of Bayesian models for spatial and temporal data and provide
practical experience of how to fit these using the WinBUGS software.
Course content: Overview of Bayesian inference and MCMC methods;
hierarchical models; autoregressive models for temporally and spatially
correlated data; Bayesian kriging and spatial prediction; measurement error
and missing data; practical experience of using WinBUGS with examples drawn
from e.g. environmental epidemiology, forest ecology, meteorological
prediction and pollution modelling.
Who should attend: Statisticians, environmental scientists, data analysts,
PhD students and postgraduate researchers working with environmental or
spatial/temporal data who wish to use a Bayesian approach to analysing their
data. No previous experience of using WinBUGS or Bayesian methods is
necessary, but participants will be expected to be familiar with standard
statistical modelling techniques such as regression and generalised linear
models; a basic understanding of hierarchical (multilevel) models will also
be useful but not essential. Participants are also strongly advised to
attend the half-day meeting prior to the course on Bayesian methods in
environmental statistics organised at the RSS headquarters in London at 2pm
on Wed 11th June 2003 (further details will be sent with registration
details).
See http://www.jiscmail.ac.uk/files/envstat/course_0306.html for further
details and registration form
___________________________
Short Courses in Modern Statistical Methods in Epidemiology
These courses are designed to give in-depth knowledge of the
principles and applications of modern statistical methods in epidemiology,
with a particular focus on Bayesian implementation.
There is a large practical component to each course with time for
hands-on data analysis. The courses are designed to be of interest to
researchers in areas such as biostatistics, epidemiology, medical geography
and environmental science, together with public health specialists,
regulators and other health-care professionals with an interest in
understanding and applying advanced quantitative methods.
Registration
Enquiries in the first instance should be directed to Wolfson
Conference Centre (Tel: +44 (0)20 8383 3117/3227/3245; email:
[log in to unmask]). Places on the course are limited and so early booking
is advisable.
Venue
All the courses will be held at the dedicated teaching facility at
the Department of Epidemiology and Public Health, Imperial College London.
Fees
The course fees are £200 per day, with an academic daily rate of
£150. A further discount of £50 per day is available to all Imperial College
staff and students.
Bayesian Hierarchical Models
3 day course September 3-5 September 2003
Instructors:
Professor Sylvia Richardson, Dr Nicky Best and Dr Clare Marshall,
Imperial College London
Summary of the course content:
* Fundamentals of Bayesian inference
* Markov chain Monte Carlo methods
* The Gibbs sampler
* MCMC in practice - its strengths and weaknesses; convergence
issues
* hierarchical models in the analysis of epidemiological data
* practical experience of using the WinBUGS software.
Spatial Epidemiology
2 day course September 8-9, 2003
Instructors:
Professor Sylvia Richardson, Dr Nicky Best and Professor Paul
Elliott, Imperial College London
Summary of the course content:
* Strengths and weaknesses of spatial studies
* Confounding
* Environmental exposures
* Ecological vs individual design
* Bayesian approaches to disease mapping
* Formulation of spatial models ( Markov-random field models,
multivariate models with spatially structured covariance)
* Hierarchical spatial models with spatially dependent random
effects
* Analysis of geo-referenced disease and exposure data using the
WinBUGS/ GeoBUGS software.
Introduction to genetic epidemiology
2-day course September 10-11,2003
Instructors:
David Balding and John Whittaker, Imperial College London
Cathryn Lewis, King's College London
This course provides an introduction to the methods used to
analyse genetic epidemiology data in order to to assess the role of
genes in causing disease, and to locate and characterize the genes
involved. The course assumes a good grasp of basic statistics
including likelihood-based estimation. No background in genetics is
required, but familiarity with the basic genetics terms would be helpful.
Our emphasis is on the principles of data analysis and study
design rather than on the details of particular computational
implementations.
Summary of course content:
* Introduction to some relevant genetics
* Segregation analysis
* Quantitative traits analysis
* Parametric linkage analysis (pedigree analysis)
* Non-parametric linkage analysis (e.g. affected sib-pairs)
* Association studies: case-control data
* Association studies: family data (e.g. TDT)
A series of case studies will be presented, and there will be
several
small computer-based exercises using S-plus and/or R.
Statistical analysis of genetic association studies
1 day course, September 12, 2003
Instructors:
David Balding and John Whittaker, Imperial College London
Andrew Morris, Wellcome Trust Centre for Human Genetics, Oxford.
The course will be at a more advanced level than the introductory
course, and will build on the ideas introduced in the last two
modules
of that course. The two courses may be taken together, or this
course
may be taken alone by researchers who already have a firm grasp of
basic statistics and some familiarity with human genetics.
Course content:
* Population genetics of linkage disequilibrium
* Statistical measures of association
* Family-based association studies
* Population-based candidate-gene studies, fine-scale mapping and
genome-wide scans
* Haplotypes vs genotype data; haplotype tagging markers
* Accounting for population substructure using case-control data
The availability of computer software implementing the methods will
be discussed,
but the course will emphasize underlying statistical principles
rather than the details of
particular computer packages.
-------------------------------------------------------------------
This list is for discussion of modelling issues and the BUGS software.
For help with crashes and error messages, first mail [log in to unmask]
To mail the BUGS list, mail to [log in to unmask]
Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html
Please do not mail attachments to the list.
To leave the BUGS list, send LEAVE BUGS to [log in to unmask]
If this fails, mail [log in to unmask], NOT the whole list
|
