Graphical Models

Graphical models

2 hours of lectures per week.
Lecturer:
Steffen L. Lauritzen, Aalborg University

Contents:
The course will give an introduction to graphical models based on directed acyclic graphs and their applications. Subjects covered include: Markov properties on directed graphs, Bayesian networks,local exact computation, graphical models for genetic analyses, the EM algorithm for Bayesian networks, Bayesian graphical models. The course will be based on a combination of material from Lauritzen (1996), Cowell et al. (1999), lecture notes, selected journal articles, as well as graphical model software.

Prerequisites
Basic understanding of probability and statistics is needed. It is an advantage to be familiar with elements of undirected graphical models, for example at the level of Lauritzen (1989).

Textbooks
S.L. Lauritzen: Lectures on Contingency Tables. 3rd ed. Department of Mathematics, Aalborg University.1989.
Electronic version 2002: \verb+www.math.auc.dk/~steffen/cont.pdf+.

S.L. Lauritzen: Graphical Models. Clarendon Press, Oxford, 1996.

R.G. Cowell, A.P. Dawid, S.L. Lauritzen and D.J. Spiegelhalter: Probabilistic Networks and Expert Systems. Springer, New York, 1999.

Software:
HUGIN Lite (\verb+www.hugin.com+). WinBUGS (\verb+www.mrc-bsu.cam.ac.uk/bugs/+.

Home page:
Further information, course material other than listed above etc., will be made available through www.math.auc.dk/~steffen/gm03/+\subsection*

ECTS-credits
10.

Semester
Autumn 2003.