geoGAM: Select Sparse Geoadditive Models for Spatial Prediction

A model building procedure to select a sparse geoadditive model from a large number of covariates. Continuous, binary and ordered categorical responses are supported. The model building is based on component wise gradient boosting with linear effects and smoothing splines. The resulting covariate set after gradient boosting is further reduced through cross validated backward selection and aggregation of factor levels. The package provides a model based bootstrap method to simulate prediction intervals for point predictions. A test data set of a soil mapping case study is provided.

Author
Madlene Nussbaum [cre, aut], Andreas Papritz [ths]
Date of publication
2016-10-29 10:48:22
Maintainer
Madlene Nussbaum <madlene.nussbaum@env.ethz.ch>
License
GPL (>= 2)
Version
0.1-1

View on CRAN

Man pages

berne
Berne - soil mapping case study
berne.grid
Berne - very small extract of prediction grid
bootstrap.geoGAM
Bootstrapped predictive distribution
geoGAM
Select sparse geoadditive model
methods.geoGAM
Methods for 'geoGAM' objects
predict.geoGAM
Prediction from fitted geoGAM model

Files in this package

geoGAM
geoGAM/NAMESPACE
geoGAM/NEWS
geoGAM/data
geoGAM/data/berne.rda
geoGAM/data/berne.grid.rda
geoGAM/data/datalist
geoGAM/R
geoGAM/R/bootstrap.geogam.R
geoGAM/R/predict.geogam.R
geoGAM/R/f.geoam.model.selection.R
geoGAM/R/methods.geogam.R
geoGAM/MD5
geoGAM/DESCRIPTION
geoGAM/man
geoGAM/man/methods.geoGAM.Rd
geoGAM/man/predict.geoGAM.Rd
geoGAM/man/bootstrap.geoGAM.Rd
geoGAM/man/berne.grid.Rd
geoGAM/man/berne.Rd
geoGAM/man/geoGAM.Rd