A guide to generalized additive models in crop science using SAS and R .

摘要

Linear models and generalized linear models are well known and are used extensively in crop science. Generalized additive models (GAMs) are less well known. GAMs extend generalized linear models through inclusion of smoothing functions of explanatory variables, e.g., spline functions, allowing the curves to bend to better describe the observed data. This article provides an introduction to GAMs in the context of crop science experiments. This is exemplified using a dataset consisting of four populations of perennial sow-thistle ( Sonchus arvensis L.), originating from two regions, for which emergence of shoots over time was compared.