2007 SSC Annual Meeting

2007 Annual Meeting of the SSC in St. John's

Generalized Additive Models

June 10th, 2007
St. John's, Newfoundland


Simon Wood (University of Bath, UK)

The Biostatistics Section is pleased to announce that Simon Wood (University of Bath, UK) will be leading a one-day workshop on generalized additive models at the 2007 Annual Meeting in St. John's, Newfoundland. Professor Wood has long-standing interests in this area, and is the author of a recent book on the topic. In addition to his general methodological interests, Professor Wood also focusses on applications of statistical modelling to ecological problems. A description of the workshop is as follows.

Generalized Additive Models (GAMs) are generalized linear models in which the relationship between a response variable and multiple predictors is described partly by a sum of smooth functions of the predictors. GAMs offers an attractive balance between flexibility and structure and since their introduction by Hastie and Tibshirani in 1986 have been widely applied across biostatistics and beyond (for example in ecology, epidemiology and finance) . Recently the GAM framework has been strengthened by the advent of reliable methods for selecting the degree of smoothness for the model components, by the development of efficient ways of representing smooth functions of several predictors, and by the integration of GAMs and mixed models. This course will cover model building and checking for GAMs within this modern framework, as well as inference with well fitting models. The focus will be practical (using the mgcv package in R) and aimed at a general statistical audience, but with examples drawn from biostatistics and ecology. Underlying theory will be presented to support the practical material (e.g. on penalized regression, the Bayesian approach to inference and smoothness selection methods), with the emphasis will be on providing a conceptual overview of this material, rather than on technical details.

Participants are encouraged to bring a laptop with a recent version of R installed.

Reference: Wood, SN, 2006, Generalized Additive Models: An Introduction with R, Chapman & Hall/ CRC, Boca Raton, Florida