INTRODUCTION TO BAYESIAN ANALYSIS USING WinBUGS

Faculty of Medicine

Imperial College London

St. Mary's Campus, London

28-30 January 2009

Tutors

Professor Nicky Best and Professor Sylvia Richardson, Imperial College London.

Course outline

This introductory course is aimed at statisticians, data analysts and quantitative researchers who are interested in finding out what Bayesian methods are all about, and how to apply them to analyse their own data using the WinBUGS software. There is a large practical component to this course with time for hands-on data analysis using examples drawn mainly from the social and health sciences.

No previous experience of Bayesian methods or WinBUGS is necessary, although familiarity with standard statistical terminology and a good grasp of the basic principles of standard (maximum likelihood-based) linear and generalised linear regression models will be assumed; participants will also be expected to be familiar with some common probability distributions (normal, binomial, Poisson).

Summary of course content

Basic probability theory (conditional probabilities, Bayes theorem)
Common probability distributions (normal, Poisson, binomial, beta, gamma)
Conjugate Bayesian inference for binary, count and continuous data
Making inference from posterior distributions
Choosing prior distributions
Introduction to Monte Carlo simulation methods, MCMC methods and Gibbs sampling
Bayesian regression models (linear, non-linear, GLMs)
Bayesian model checking and comparison (residuals, predictions, DIC)
Random effects (multilevel / hierarchical) models for normal data
Hands-on experience of using WinBUGS software for Bayesian analysis
Opportunity to discuss analysis of your own data with the course leaders and other participants

Location

The course will take place in the Division of Epidemiology, Public Health and Primary Care of the Faculty of Medicine and the participants will be able to use these facilities. Note that this is in the Faculty of Medicine, Imperial College London at St. Mary's Campus, and NOT in the main campus in South Kensington.

The complete address is

Faculty of Medicine
Imperial College London
St. Mary's Campus, Norfolk Place
W2 1PG London - UK

More information on how to arrive can be found at http://www1.imperial.ac.uk/medicine/contacts/campuses/stmarys/

Laptops and Software download

Participants are asked to bring their own laptops for the practicals if possible. We ask that you work in pairs during the practical classes, so if you do not have a laptop we can always pair you up with someone that does.

If you intend to bring your own laptop, it would save time if you could download and install the software beforehand. You will need to install WinBUGS version 1.4.3, which can be downloaded from the WinBUGS website:

http://www.mrc-bsu.cam.ac.uk/bugs/winbugs/contents.shtml

which also contains instructions for installation. You will need to download and install the main WinBUGS 1.4 package, and then the patch to upgrade this to 1.4.3.

Please also remember to register on http://www.mrc-bsu.cam.ac.uk/bugs/winbugs/register.shtml and install the key for full use, otherwise some of the practical exercises will not work.

If you are coming from outside the UK, please remember to bring an adaptor plug so that you can plug your laptop into UK power sockets!

Analysis of your own data

During the short course: Time will be set aside during the course for each participant to give a short presentation about a problem that they are interested in using Bayesian methods to solve, with feedback from course tutors on how to achieve this. However, it is not necessary to bring your own data, as it is not intended for participants to actually analyze their own data during the course.

Phase II mentoring scheme (limited places available): Participants who would like further help with analyzing their own data using WinBUGS are invited to apply to the Phase II mentoring scheme. The mentoring scheme is designed to provide you with one-to-one help with developing your model, implementing your analysis in WinBUGS and interpreting the output and results. The overall goal will be to produce a journal paper (or equivalent) based on the analysis, which is jointly authored by the participant (and colleagues, as appropriate) and the mentor(s) from the BIAS II project team.

If you are interested in applying for this scheme, please tick the appropriate box on the registration form. We will then contact you separately, asking you to provide a short description of the dataset/problem than you would like help with analyzing, including reasons why you think a Bayesian analysis might be appropriate. Please note that places on the mentoring scheme are limited, and we will select applicants on the basis of which projects we judge to be most suitable for analysis using WinBUGS. Selected applicants will also be required to provide a letter of support from their supervisor/line manager/head of department. Further details of the mentoring scheme will be provided during the short course, and applicants will not be selected until after the course. Note that there is no additional charge for participating in the mentoring scheme.

Course fees and Registration

£585 (standard fee)
£150 (reduced fee for participants from academic and other non-profit research organisations)
£75 (reduced fee for PhD students)

The registration form for the short course can be found at: http://www5.imperial.ac.uk/medicine/coursebookings/introbayesianwinbugs_0901.html

There is no closing date for registration, but places on the course are limited and tend to get booked very quickly.

Course Materials

Click to download the course materials

BayesIntro.zip

data.zip