Predict with the final model

predict() and augment() on a nested_final_fit call the trained workflow's own methods. Each returns what the same call on the workflow from extract_workflow() returns.

Usage

# S3 method for class 'nested_final_fit'
predict(object, new_data, type = NULL, opts = list(), ...)

# S3 method for class 'nested_final_fit'
augment(x, new_data, eval_time = NULL, ...)

Arguments

object, x

A nested_final_fit from nested_final_fit().

new_data

A data frame of new observations to predict.

type, opts

Passed to workflows::predict.workflow() unchanged. type selects the prediction type, such as "numeric", "prob" or "survival". Left unset, it takes the workflow's default.

...

For predict(), arguments for the model's predict method, such as level with type = "conf_int". parsnip gives an error for a name outside its list of predict arguments. For augment(), not used, and must be empty.

eval_time

For censored regression, the times at which to evaluate survival probabilities in augment(). Ignored otherwise.

Value

predict() returns the workflow's tibble of predictions. augment() returns the prediction columns, then the columns of new_data, with a .resid column for a regression model when the outcome is present.

Details

Residuals on the rows the model was fit on are in-sample, so they do not measure performance. Report the nested estimate from collect_metrics() on the nested run the fit was built from.

See also

nested_final_fit(), extract_workflow(), workflows::predict.workflow(), workflows::augment.workflow()

Examples

data(mtcars)

rec <- recipes::recipe(mpg ~ ., data = mtcars) |>
  recipes::step_pca(recipes::all_predictors(), num_comp = tune::tune())
wf <- workflows::workflow(rec, parsnip::linear_reg())

set.seed(1)
folds <- nested_resamples(
  mtcars,
  outside = rsample::vfold_cv(v = 2),
  inside = rsample::vfold_cv(v = 2)
)
set.seed(2)
res <- nested_tune_grid(wf, folds, grid = data.frame(num_comp = 1:2))
set.seed(3)
final <- nested_final_fit(wf, res)
predict(final, new_data = mtcars[1:3, ])
#> # A tibble: 3 × 1
#>   .pred
#>   <dbl>
#> 1  23.1
#> 2  23.1
#> 3  25.2
predict(final, new_data = mtcars[1:3, ], type = "conf_int", level = 0.9)
#> # A tibble: 3 × 2
#>   .pred_lower .pred_upper
#>         <dbl>       <dbl>
#> 1        22.1        24.2
#> 2        22.1        24.2
#> 3        23.9        26.5
augment(final, new_data = mtcars[1:3, ])
#> # A tibble: 3 × 13
#>   .pred .resid   mpg   cyl  disp    hp  drat    wt  qsec    vs    am
#> * <dbl>  <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1  23.1  -2.14  21       6   160   110  3.9   2.62  16.5     0     1
#> 2  23.1  -2.14  21       6   160   110  3.9   2.88  17.0     0     1
#> 3  25.2  -2.38  22.8     4   108    93  3.85  2.32  18.6     1     1
#> # ℹ 2 more variables: gear <dbl>, carb <dbl>