[R-pkgs] caret version 6.0-21

[Kuhn, Max]

The caret package (short for Classification And REgression
Training) attempts to streamline the process for creating
predictive models. The package contains tools for:

 - data splitting
 - pre-processing
 - feature selection
 - model tuning using resampling (with parallel processing)
 - variable importance estimation

as well as other functionality. The package website has lengthy
descriptions of functionality:

   http://caret.r-forge.r-project.org/

A new version, 6.0-21, was just released to CRAN and contains
major improvements:

 - A new system for user-defined models has been added. See
  http://caret.r-forge.r-project.org/custom_models.html.

 - When creating the grid of tuning parameter values, the column
  names no longer need to be preceded by a period. Periods can
  still be used as before but are not required.

 - trainControl now has a 'method = "none"' resampling option that
  bypasses model tuning and fits the model to the entire training
  set.

- Several new models were added, bringing the total number of
  models up to 149.

 - ggplot methods were added for several classes.

Most changes are not user-visible so that the impact on current
scripts should be minimal.

For this caret update, the AppliedPredictiveModeling package will be
brought up to date shortly with slightly modified scripts for the
analyses contained in the text.

Thanks to the R community and caret contributors: Jed Wing, Steve
Weston, Andre Williams, Chris Keefer, Allan Engelhardt, Tony Cooper,
Zachary Mayer and the R Core Team.

Please contact me (max.kuhn at pfizer.com) with any comments, suggestions,
questions or ideas for a good wedding toasts.

Max