- Added a workaround which allows not providing the unused variables when the terms argument is used.
- Added a
meta_only
option toplot.metagam()
. - Bug fix in plots.
- Added back label option for plot_dominance.
- A large number of internal changes which make package maintenance easier and the package more stable.
- Standard error of intercept is now included in simulations for p-values and simultaneous confidence intervals when type="iterms".
- 'knitr' moved from Imports to Suggests.
- Fixed a bug which caused metagam to crash when there were special characters (in particular '_' and '.') in variable names.
- This is a major revision of the package containing the changes described since 0.2.0 below.
- Large number of dependencies removed, to make the metagam less sensitive to errors in other packages and to make it easier to install on different platforms.
- Simultaneous confidence intervals have been added.
- Computation of p-values has been completely revised. The previous method would give nonsensical answers in cases where the effects pointed in different directions in different datasets.
- A vignette on p-values and simultaneous confidence intervals has been added.
- removed viridis from Suggests, to reduce the number of dependencies.
- removed gratia from Suggests.
- Parallelism support removed, because it was buggy. Will refactor the code before considering adding it back.
- Proper support for parallelism added, thanks to Henrik Bengtsson. See example in help("metagam").
- Fixed issue #30. Running in parallel caused error.
- Updated citation after paper was published in NeuroImage.
- Updated recommended citation after paper was accepted in NeuroImage.
- README is updated to suggest installation using Bioconductor
- A message is now printed to the user on startup if multtest is not installed. As a consequence, multtest has been added to Suggests:
- Moved package 'viridis' from Imports to Suggests, since 'viridis' is hard to install on some platforms.
- Internal change which makes metagam compatible with dplyr 1.0.0.
- This is the initial release.