Title: Statistical and ecological modeling in AD Model Builder
Date: August 5, 2006 (the day before the Joint Statistical Meetings open),
Time: 9am to 4pm
Venue: Room 203, School of Aquatic and Fisheries Building, University
of Washington, Seattle
Instructors: Mark Maunder and Hans Skaug
Assistant: Eric Ward
Cost: Free
This course targets statisticians and ecologists who want to learn AD
Model Builder.
AD Model Builder (ADMB http://otter-rsch.com/admodel.htm) is a
commercially available
software programming language that aids the development of complex
statistical models.
With its recently developed random effects module
(http://otter-rsch.com/admbre/admbre.html)
ADMB now also support non-linear mixed models.
The strengths of ADMB are:
1. Flexibility in model formulation (based on C++ like language).
2. Fast and stable numerical routines (based on Automatic Differentiation).
3. Suitable for both Bayesian and frequentist frameworks
4. A large example suite is available
(http://otter-rsch.com/admbre/examples.html).
Who will this course be useful for?
1. Those who know about ADMB, and want to learn more, particularly
about the new random effects module.
2. Those with complex models, possibly involving random effects, who
are looking for a computational framework for solving the problem.
Simple example applications using ADMB
- Treating the probability of detection as a random effect in
mark-recapture models
- Integrating a population dynamics model with a GLM of survey data
- Flexible nonlinear mixed models for various fields
- GLM's with spatial error structure
Use this opportunity to learn ADMB while you are at the Joint
Statistical Meetings!
A PC based laptop computer is required.
For more information contact
Mark Maunder
Inter-American Tropical Tuna Commission
Tel: (858) 546-7027
[log in to unmask]
or
Hans Skaug
University of Bergen
[log in to unmask]
|