Dina Research School


Monte Carlo methods for Hierarchical (Mixed) Models.

Koldkærgård Landboskole, Skejby, October 30-31 2003.

Thursday, October 30

11.00 Arrival and accommodation.
12.00 Lunch
13.00 Introduction and presentation of participants
Erik Jørgensen , Dina Research School.
13.15 Theory session I:
Rasmus Waagepetersen , Aalborg University.
The concepts of hierarchical and generalized linear mixed models are discussed in relation to various datasets from the agricultural sciences. A basic problem with these models is that the likelihood function is not available in closed form. We consider various numerical techniques for calculating the likelihood function with a particular focus on Monte Carlo methods.
14.00Short break followed by computer exercises.
14.45Discussion of exercises
15.00 Coffee break.
15.30Theory session II:
Rasmus Waagepetersen , Aalborg University.
We continue the discussion of Monte Carlo methods and consider importance sampling and Markov chain Monte Carlo methods for generating samples from an importance sampling distribution.
16.15 Computer exercises.
17.00 Theory session III:
Rasmus Waagepetersen , Aalborg University.
We discuss the Bayesian approach to inference for hierarchical models and how it can be implemented using the software BUGS.
18.00 Dinner.
19.00 Case study I: .
19.45 Computer exercises (continued) .
21.45 Coffee and sandwich.
 

Friday, October 31

7.30 Breakfast.
8.30 Discussion of computer exercises .
9.00 Case II:
Ole Fredslund Christensen , Center for Bioinformatics, Aarhus Universitet .
9.45 Coffee Break.
10.00Theory IV: .
Rasmus Waagepetersen , Aalborg University.
The final lecture is devoted to more specialized topics in MCMC and hierarchical models with non-Gaussian random effects.
10.30 Break.
10.35 Computer exercises.
11.15 Case III:
11.45 Discussion and Closing .
Erik Jørgensen , Dina Research School.
12.00 Lunch and departure.
Dina logoAuthor: phd@dina.kvl.dk. Updated: 28 april 2001