Wrappers for extracting feature importances from a fitted model using common modeling backends

`interpret_caret`, `interpret_tidymodels`, and `interpret_h2o` are wrappers for extracting feature importances from a fitted model using caret, tidymodels, and h2o backends, respectively. These wrapper functions provide uniformity of input arguments to easily switch between the different modeling packages (see `interpret_models()`).

interpret_caret(fit, ...)

interpret_h2o(fit, ...)

interpret_tidymodels(fit, ...)

Arguments

fit

Model fit. Typically the output of `fit_caret()`, `fit_h2o()`, or `fit_tidymodels()`.

...

Additional arguments to pass to `varImp()` in `fit_caret()`, `h2o.varimp()` in `fit_h2o()`, or `vip::vi()` in `fit_tidymodels()`.

Value

A tibble with the following columns:

Variable

Name of feature/variable.

Importance

Feature importance score.

See also

Other interpret_models_family: interpret_models()