Adding support for new model classes

Adding (almost) full support to the easystats ecosystem

To add support for models in the easystats ecosystem, especially for the packages parameters, or to support most / all features from easystats, following methods should be added to the insight package:

  • insight::find_formula()
  • insight::model_info()
  • insight::get_parameters()
  • insight::get_statistic()
  • insight::find_statistic()
  • insight::get_varcov()
  • insight::get_df()

Nice to have (though some of the following functions could already work if the above methods are implemented):

  • insight::find_parameters()
  • insight::get_modelmatrix()
  • insight::get_predicted()
  • insight::link_inverse() and insight::link_function()
  • insight::get_residuals()

Adding basic support

If the purpose is just “tidy” output, it is enough to add following methods to the parameters package:

  • parameters::ci()
  • parameters::p_value()
  • parameters::standard_error() or insight::get_varcov()

The simplest way is just adding a method for parameters::model_parameters(), however, than this model-class is not supported by all functions we offer in the easystats ecosystem.

See also https://easystats.github.io/effectsize/articles/effectsize_API.html