Generalized Additive Models
By Simon N. Wood. 2006
9,043 citations recorded by the source corpus.
Generalized additive models are generalized linear models in which the linear predictor includes a sum of smooth functions of covariates, where the shape of the functions is to be estimated. They have also been generalized beyond the original generalized linear model setting to distributions outside the exponential family and to situations in which multiple parameters of the response distribution may depend on sums of smooth functions of covariates. The widely used computational and inferential framework in which the smooth terms are represented as latent Gaussian processes, splines, or Gaussian random effects is reviewed, paying particular attention to the case in which computational and theoretical tractability is obtained by prior rank reduction of the model terms. An empirical Bayes approach is taken, and its relatively good frequentist performance discussed, along with some more overtly frequentist approaches to model selection. Estimation of the degree of smoothness of component functions via cross validation or marginal likelihood is covered, alongside the computational strategies required in practice, including when data and models are reasonably large. It is briefly shown how the framework extends easily to location-scale modeling, and, with more effort, to techniques such as quantile regression. Also covered are the main classes of smooths of multiple covariates that may be included in models: isotropic splines and tensor product smooth interaction terms.
Subfields: Management Science and Operations Research
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Management Science and Operations Research
- Jan Peters
- Yuedong Wang
- Dominik Janzing
- Bernhard Schölkopf
- Alan F. Karr
- Xiaodong Lin
- Martijn B. Wieling
- Clark Glymour
Citations in the corpus
- Generalized Sparse Additive Models (2019, methodology): “This model is known as a generalized additive model (GAM) [Hastie and Tibshirani, 1990].”
- Spatial and temporal variability in somatic growth of green sea turtles (Chelonia mydas) resident in the Hawaiian Archipelago (2004, methodology): “The models were fitted here using the gam() function in SPLUS (Hastie 1993), although these same models can be readily fitted using the vgam() supplementary library (Yee and Wild 1996) in R (Ihaka and Gentleman 1996).”
- Impacts of agricultural intensification on bird communities: New insights from a multi-level and multi-facet approach of biodiversity (2016, methodology): “Then, to determine the importance of fixed effects while limiting uncertainty from model selection, a model averaging procedure was applied (Wood, 2006) (R package {MuMIn}; Barton, 2015) on the models fitted with the Maximum Likelihood method (which allows comparing fixed-effects structure of mixed…”
- Fast growing and interpretable oblique trees via logistic regression models (2009, methodology): “IC = −2l(θ̃PL(λ)) + cp× edf where l and θ̃PL are as before, cp is some complexity parameter and edf is a measure of the effective degrees of freedom [Hastie and Tibshirani, 1990] of the penalized model.”
- Synoptic and meteorological drivers of extreme ozone concentrations over Europe (2016, background): “It includes a static non-linear transformation (link-function) and the response is not restricted to a normal distribution (Wood 2006).”
- Interpretability is Harder in the Multiclass Setting: Axiomatic Interpretability for Multiclass Additive Models (2018, methodology): “Generalize additive models (GAMs) were first introduced to model features with more flexible base learners (Hastie and Tibshirani 1990; Wood 2006).”
- A comparison of methods for the fitting of generalized additive models (2008, methodology): “Various techniques can be employed for actually fitting such models, some of them documented in two recent monographs (Ruppert et al. 2003; Wood 2006), and there is ongoing research on new ones.”
- How to Make Models Add Up — A Primer on GLMMs (2009, methodology): “Random effects can also be added, to form GAMMs, Generalized Additive Mixed Models (Wood 2006).”
- Penalized Partial Least Squares with Applications to B-Spline Transformations and Functional Data (2006, background): “, fp are represented in terms of basis functions [11].”
- Evaluation of spatial predictions of site index obtained by parametric and nonparametric methods—A case study of lodgepole pine productivity (2005, methodology): “An iterative algorithm was used for computing all functions (Hastie and Tibshirani, 1990).”
- Rise of the generalists: evidence for climate driven homogenization in avian communities (2012, methodology): “The GAMs were constructed using version 1.6–1 of the mgcv package (Wood, 2006) in 196 the statistical package R (R Core Development Team 2009).”
- Modelling sperm whale habitat preference: a novel approach combining transect and follow data (2011, methodology): “A Generalized Additive Model (GAM) framework was used to model the relationship be - tween sperm whale presence/absence at each GPS fix and the predictors described above (Hastie & Tibshirani 1990, Wood 2006).”
- Smooth Design-Adapted Wavelets for Nonparametric Stochastic Regression (2004, methodology): “1 Measurement of exhaust from burning ethanol The first data set, which has already been discussed by many authors [7, 10, 24, 26], consists of 88 measurements from an experiment in which ethanol was burned in a one-cylinder automobile engine.”