stacks 0.2.2

• Fixed errors arising from outcome levels that are not valid column
  names in the multinomial classification setting.
• Fixed collect_parameters failing to return stacking coefficients
  in the two-class classification setting.
• Regenerated example objects with updated {rsample} fingerprinting--saved
  model objects may need to be regenerated in order to build stacks combining
  models generated before and after this update.
• Various other bug fixes and improvements to documentation.

stacks 0.2.1

• Updates for importing workflow sets that use the add_variables() preprocessor.
• Plot fixes for cases where coefficients are negative.
• Performance and member plots now show the effect of multiple mixture values.
• Package diagrams now have alt text.
• Various bug fixes and improvements to documentation.

stacks 0.2.0

This release integrates the package more closely with the tidymodels ecosystem, including support for finetune and workflowsets. It also introduces support for fitting ensembles with an elastic net and features a number of bug fixes and other improvements.

Breaking changes

This release of the package changes some elements of the internal structure
of model stacks. As such, model stacks stored as saved objects will need to
be regenerated before predicting, plotting, printing, etc.

New features

• The package now supports elastic net models as a meta-learner via
  the mixture argument to blend_predictions.
• The package can now add candidates from workflow_map objects
  from the new {workflowsets} package. The interface to add_candidates
  for doing so is the same as with tune_results objects, and
  add_candidates is now a generic function.
• Objects tuned with racing methods from the {finetune} package can now be
  added as candidate members.

Bug fixes

• Fixed bug in determining member hyperparameters during member
  fitting when using non-RMSE/ROC AUC metrics.
• Fixed bug arising from model definition names that are not valid column
  names. The package will now message in the case that the provided names
  are not valid column names and use make.names for associated candidate
  members.

Miscellaneous improvements

• Drop {digest} dependency in favor of {tune}/{rsample} "fingerprinting"
  to check consistency of resamples.
• fit_members() will now warn when supplied a model stack whose
  members have already been fitted.
• Integrate with {tune} functionality for appropriately coloring errors,
  warnings, and messages.
• Improved faceting and axis scales to make autoplot with type = "members"
  more informative.
• Various improvements to documentation.

stacks 0.1.0

Initial release on CRAN!