Short course on Hierarchical Generalized Linear Models -10 May 2006
- Prof Roger Payne, Rothamsted Research & VSN International
- 14:00 - 17:00, Wed 10th May 2006
- Room C.18, Ferranti Building, Sackville Street Campus, University of Manchester
Introduction
Hierarchical Generalized Linear Models (HGLMs) provide a flexible and efficient framework for modelling non-Normal data in situations when there may be several sources of error variation. They are defined by extending the familiar generalized linear models (GLMs) to include additional random terms in the linear predictor. They include generalized linear mixed models (GLMMs) as a special case, but do not constrain the additional terms to follow a Normal distribution and to have an identity link (as in the GLMM). For example, if the basic generalized linear model is a log-linear model (Poisson distribution and log link), a more appropriate assumption for the additional random terms might be a gamma distribution and a log link. HGLMs thus bring a wide range of models together within a single framework. Each HGLM is made up from two interlinked generalized linear models, so we have access to a familiar repertoire of model checking techniques to help determine the appropriate error distributions and models.
This course will introduce the underlying theory and show examples of situations where HGLMs can be useful. It will use the GenStat procedures that have been written by Lee, Nelder & Payne to implement the methodology, and which will accompany the forthcoming book on HGLMs by Lee, Nelder & Pawitan (due in mid 2006).
References
- Lee, Y. & Nelder, J.A. (1996). Hierarchical generalized linear models (with discussion). J. R. Statist. Soc. B, 58, 619-678.
- Lee, Y. & Nelder, J.A. (2001). Hierarchical generalized linear models: a synthesis of generalised linear models, random-effect models and structured dispersions. Biometrika, 88, 987-1006.
- Lee, Y. & Nelder, J.A. (2006). Double hierarchical generalized linear models (with discussion). Appl. Statist., 55, 1-29.
Target Audiences and Registration
The course is free of charge. It is mainly for postgraduate students in Statistics or Applied Statistics, but anyone who is interested is welcome to attend. Please send an email to Jianxin.Pan[at] manchester.ac.uk. The maximum number of audiences we can take is about 40. Please note it is first come first serviced.
Timetable
To be posted here.
Directions
The Ferranti building is number 20 on the University Campus map, and lies within 5 minutes walk of Piccadilly railway station, which hosts mainline services to most areas of the UK (London, for example, is 2 hours away, and Manchester International Airport about 20 minutes). Detailed directions are available on the general information pages.