- User defines the distance metric to be used in the interpolation process
- Genetic distance uncertainty can be used to build the variogram
- Overall performance improvement
- New vignette describing new features in this version
- added functions 'print.gv' and 'summary.gv' for fast view of gv object data
- Improved variogram model fitting in function gv.model (and related functions mtest.gv and predict.gv)
- Added new models (pentaspherical and cubic)
- Improved the calculation of the linear model with sill
- Improved the nls fitting
- Improve distance calculations with resistance/friction information using 'gdistance' package.
- The "krig" function was modified to accept a distance function to calculate distances. The 'default' uses simple euclidean distances between coordinates to maintain compatibility with code for previous versions.
- Plot method for 'gv' was duplicated and is now corrected.
- Plotting gv with multiple trees now displays CI bars instead of lines.
- Bug corrected: plot.gv wasn't displaying the variogram correctly if NA were present.
- Explicit declaration of imports in NAMESPACE to avoid a NOTE during CRAN check.
- Variograms can now be generated with the full (or part of) tree posterior probability instead of a single consensus tree.
- Added a generic function to generate the variogram if a single or multiple trees are found.
- Added new variogram plot to deal with multiple trees.
- If the variogram produces NAs, a warning is thrown.
- Krig function does a check to verify if grid coordinates are given as a 'data.frame'.
- Some minor fixes to correct the NOTES on CRAN checks (register S3 classes on NAMESPACE)
- The tutorial as added as a vignette.
- First version submitted to CRAN