Changes in version 0.2.2 (2022-06-18)                  

  - Added class structure for accessor function objects
  - Ensured consistency in documentation.
  - Added new function, heterogeneity_CLAN(), that investigates the
    presence of treatment effect heterogeneity along all CLAN variables.
  - Added function get_best() that returns the best learner.
  - Changed behavior of get_CLAN() to not plot ATE estimates when plot =
    TRUE.

                 Changes in version 0.2.1 (2022-05-12)                  

  - Replaced isa() with inherits() to avoid reliance on R >= 4.1.
  - Changed default in parallel argument in GenericML to FALSE.

                 Changes in version 0.2.0 (2022-05-06)                  

  - Replaced 1:length(x)-like loops with safer seq()-based counterparts.
  - Replaced if() conditions comparing class() to string with the safer
    isa().
  - Parallel computing is now also supported on Windows.
  - Added a method setup_plot() that returns the data frame that is used
    for plotting. Also, made the addition of ATEs in plots optional via
    the argument ATE in plot.GenericML().
  - Added a function GenericML_combine, which combines multiple
    GenericML objects into one.
  - Implemented stratified sampling for sample splitting.

                 Changes in version 0.1.1 (2021-12-07)                  

  - Fixed a few typos in the documentation.
  - Added conditions so that learners based on the package glmnet in the
    tests and examples will be skipped on Solaris machines. Note that
    this does not prevent an error on Solaris because glmnet is still a
    Suggest of GenericML and glmnet v4.1.3 cannot be reliably installed
    on Solaris machines.

                 Changes in version 0.1.0 (2021-11-24)                  

  - Initial release on CRAN (Nov. 24, 2021)