version 0.2.9
- Added a new impute_all option to linear and tree model imputation functions,
  to support predictive mean matching imputation.
- Fixed predictive mean matching imputation. (Thanks to Susie Jentoft for
  reporting.) Note that predictive mean matching can now only be used with
  predictors that support the impute_all option.
- Fixed Rd bug (thanks to Kurt Hornik for pointing out the changes).
- Fixed bug in random hot deck (Thanks to GH user sfhalapour for 
  the PR).

version 0.2.8
- Package now depends on R >= 4.0.0
- Examples using |> now in stead of magrittr::%>%
- fixed bug in impute_knn: VIM backend was ignored. (Thanks to Sander
  Scholtus and Sigrid van Hoek for reporting)

version 0.2.7
- Updated C-code for compatability with CLANG compiler (Thanks
  to Brian Ripley for pointing out the specific noncompliance 
  with WRE)

version 0.2.6
- Fixed error in 'impute_shd' caused by VIM update.

version 0.2.5
- CRAN-requested update, avoiding misuse of 'order' on data.frames
  (Thanks to Kurt Hornik for reporting).

version 0.2.4
- Updated tests in preparation of the new 'stringsAsFactors=FALSE' policy 
  (thanks to Kurt Hornik)
- Fixed parsing error for formulas with very long variable names (Thanks to 
  Karl Dunkle Werner).
- Fix: ... was not parsed to 'missForest' in 'impute_mf'.
- Fix: Column names with spaces were not handled properly (Thanks to Miles
  McBain)
- Fix: 'impute_proxy' would crash when imputing a character constant.

version 0.2.3
- switched to 'tinytest' testing framework
- Fix in impute_median: wrong medians imputed at randomly sorted data (thanks to Ranjeed-S).

version 0.2.2
- new functions 'impute' and 'impute_' use pre-fit models to impute datasets. (Thanks to Max Kuhn for suggesting this).
- new utility function 'foretell'. Replacement for 'predict' that always returns values of the predicted type.
- bugfix: ... arguments were not passed correctly for linear models.

version 0.2.1
- Added support for ridge/elasticnet/lasso regression (using 'glmnet')
- Added support for missForest imputation (using 'missForest')
- Added support for imputation using EM-estimated parameters (using 'norm')
- 'impute_proxy' now supports transformations in the rhs of formula.
- 'impute_median' gains 'type' argument to specify algorithm for median computation.
- New functions 'simputation_capabilities' and 'simputation_suggests'
  detect which packages supporting imputation methods are available.
- Renamed 'na.action' -> 'na_action' for consistency within simputation.
- Default 'na_action' for 'impute_cart' is now 'na.rpart'
- Default 'na_action' for 'impute_rf' is not 'na.roughfix'
- Complete overhaul of function documentation. Now set up systematically
  with synopsis of methodology where relevant.
- Dependencies on larger packages (randomForest, glmnet) moved to 'Suggests'
- Fix in 'impute_rf': needed workaround for bug in randomForest 4.6.12 
  (formulas like Y ~ . - X not handled correctly by predict.randomForest)
- Bugfix in 'impute_pmm': argument 'predictor' was ignored.
- Bugfix in 'impute_rhd': attributes of 'dat' were not always preserved.

version 0.2.0
- Added na.action option for model-based imputation (thanks to Soogeun Park)
- impute_rhd, impute_shd, and impute_knn now have option 'backend="VIM"'
- removed call to structure(NULL,*) to comply to changes in R-devel (thanks to Martin Maechler)

version 0.1.0
- initial release