Short Course: BAYESIAN ANALYSIS for INDUSTRY School of Computing, Engineering and Technology. University of Sunderland. UK 13 - 17 September 1999 One week full time. This new short course is supported by the Engineering and Physical Sciences Research Council Integrated Graduate Development Scheme. The short but intensive course is intended to introduce the delegate to the ideas of modern Bayesian inference and decision making under conditions of uncertainty. Successful completion of the module will attract 15 credits at level M. (180 level-M credits are required for the degree of MSc). Each delegate will receive a certificate of attendance. The course will be presented at the University's St. Peter's Campus, located on the riverside, next to St.Peter's Church (AD 674) and the National Glass Centre and a short distance from the seafront. Course Content: *Introduction to subjective probability. *Bayesian inference: simple cases with small numbers of parameters. *Introduction to Bayesian computation and software. *Conditional independence graphs and graphical modelling. *More complex cases. Markov chain Monte Carlo methods. The Gibbs sampler. *Use of suitable software. *Decision analysis. Utility. Decision trees. Influence diagrams. *Use of suitable software. *Introduction to probabilistic reasoning in expert systems. *Industrial and commercial case studies. Teaching Methods and Course Materials: Delegates will be given a reading list before attendance at the module which will introduce them to the ideas of the Bayesian paradigm. After attending the module they will be given supportive reading matter, including case studies, leading towards their module assessment. Formal class contact will consist of sessions of two types: classroom-based sessions and practical workshops. Each will be used both for theoretical knowledge and practical skills. Specially written handout material will be used to support the module in addition to guided reading of text books. The module will make use of our experience of real industrial and commercial problems which has been accumulated through research and consultancy work and in connection with our M.Sc. course in decision support systems. Final assessment will be by means of a "mini-project" based on the analysis of a practical problem in inference or decision making under uncertainty. Where possible this problem will be drawn from within the delegate's own organisation. Course Prerequisites: A delegate attending the course will be expected to have an honours degree in a numerate discipline. In particular the following elements will normally be required: 1. Familiarity with calculus up to and including an understanding of the concepts of integration with respect to more than one variable. 2. Familiarity with linear algebra up to and including an understanding of the concepts of inverses and the solution of systems of linear equations. 3. Understanding of the basic ideas of probability, for example addition and multiplication of probabilities and the normal and binomial distributions. 4. Some previous use of computers. Course Fee: The cost of the course will be £500 per student. The fee covers participation in the course, course materials, incidental tea/coffee and reception, lunches during attendance for five days at Sunderland, but does not cover travel costs, accommodation or evening meals. The full fee is payable upon notification of acceptance for the course. Accommodation: A list of hotels located in the vicinity of the school can be supplied on request. Applications: Applicants should contact as soon as possible: Diana Webb School of Computing, Engineering and Technology The Informatics Centre St Peter's Way Sunderland SR6 0DD Telephone: +44 (0)191 515 3291 Fax: +44 (0)191 515 2781 e-mail: [log in to unmask] http://osiris.sunderland.ac.uk/~cs0mfa/bafi.html %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%