Tutorial: Graphical Models and Bayesian Networks with R
University of Oslo, Norway, November 2012
Goal
Introduce participants to using R for working with graphical
models. Focus will be on graphical log-linear models for discrete data
(contingency tables) and on probability propagation in Bayesian
networks.
Teacher
Søren Højsgaard, Department of Mathematical Sciences, Aalborg
University, Denmark
When and where
The tutorial will take place on Monday, November 26 and on Tuesday, November 27.
The tutorial consists of a theoretical and practical part.
Venue: Norwegian Computing Center, Gaustadalléen 23a/b, 4. etg, Room
Alfa-Omega.
Time schedule:
Theoretical part (for all): Monday 9-12.
Practical part: (group1): Monday 13-16; (group2): Tuesday 9-12
Outline
There will be a running example about building a probabilistic expert
system for a medical diagnosis from real-world data.
- Probability propagation with Bayesian networks (BNs) and their
implementation in the gRain (gRaphical independence networks)
package.
- A look under the hood of BNs to understand mechanisms of
probability propagation.
- Dependency graphs and conditional independence restrictions.
- Log-linear models, graphical models, decompsable models and their
implementation in the gRim (gRaphical independence models) package.
- Model selection with gRim
- Converting a decompsable graphical model to a Bayesian network.
Prerequisites and preparation
A working knowledge of R is a prerequisite for participating in the
practical part of the tutorial.
There will be some R packages to be installed, and it is highly recommended to do so before
the tutorial.
The packages you MUST install are gRbase, gRain and
gRim.
When doing so, please follow the instructions given
here.
In addition, please also install the rpart package.
Course material
Signing up for the tutorial
Registration is now closed!
File translated from
TEX
by
TTH,
version 4.12.