Skip to contents

Changelog

Source: NEWS.md

gdm 1.6.0-3

CRAN release: 2024-11-01

bug fixes

  • Fix bug in gdm.predict and gdm.transform functions (also impacted gdm.crossvalidation results)

gdm 1.6.0-2

CRAN release: 2024-09-27

Updates

  • Address new CRAN C++11 requirements.

gdm 1.6.0

  • Transition to terra functionality for much faster processing of rasters.
  • Fix bug in the formatsitepair function for bio data provided as distance matrix.

gdm 1.5.0-9

bug fixes

  • Fix bug in the varImp function related to NULL models

gdm 1.5.0-8

bug fixes

  • Fix bug in the formatsitepair function when using custom weights.

gdm 1.5.0-7

bug fixes

  • Fix bug in the varImp function when predSelect = F.

gdm 1.5.0-6

bug fixes

  • Fix bug in the varImp function related to testing models with few (<4) variables.

gdm 1.5.0-5

bug fixes

  • Fix bug in the varImp function if sum of I-spline coeffs = 0 for geographic distance variable.

gdm 1.5.0-4

bug fixes

  • Fix minor bug in the varImp function.

gdm 1.5.0-3

CRAN release: 2022-04-05

  • New, much faster version of the varImp function.

bug fixes

  • Fix bug that broke varImp if using non-default number of I-splines.

gdm 1.5.0-2

  • Added verbose argument to the formatsitepair function to control whether site-pair attributes are printed.

bug fixes

  • Fix bug that prevented printing of I-spline coefficients when using the summary function.
  • Fix bug that prevented the sampleSites argument in the sitepairformat function from actually subsampling the site-pair table.

gdm 1.5.0-1

CRAN release: 2021-11-18

bug fixes

  • Fix bug associated with the formatting of the gdm model object that broke the predict.gdm function.

gdm 1.5.0

CRAN release: 2021-11-16

  • Substantial update to the package structure to use devtools, roxygen2, etc. development tools.

  • Fixed a number of bugs associated with using non-default numebrs of splines.

  • Add two new functions gdm.partition.deviance and gdm.crossvalidation. The gdm.partition.deviance functions partitions deviance explained by different sets of predictors into unique/shared components. The gdm.crossvalidation function produces a number of model evaluation metrics using bootstrapping.

  • summary now provides the sum of the I-spline coefficients (indicator of predictor importance) and lists variables from highest sum to lowest.

  • plot now plots I-splines in order from highest sum of coefficients to least.