tests/testthat/_snaps/prob-roc_auc.md

In topepo/yardstick: Tidy Characterizations of Model Performance

roc_auc() hand-till method ignores levels with 0 observations with a warning (#123)

Code
  expect_identical(roc_auc_vec(truth, estimate), 0.5)
Condition
  Warning:
  x No observations were detected in `truth` for level: y.
  i Computation will proceed by ignoring those levels.

warning is thrown when missing events

Code
  out <- roc_auc(no_event, truth, Class1)[[".estimate"]]
Condition
  Warning:
  No event observations were detected in `truth` with event level 'Class1'.

warning is thrown when missing controls

Code
  out <- roc_auc(no_control, truth, Class1)[[".estimate"]]
Condition
  Warning:
  No control observations were detected in `truth` with control level 'Class2'.

multiclass one-vs-all approach results in multiple warnings

Code
  out <- roc_auc(no_event, truth, Class1, Class2, estimator = "macro")[[
    ".estimate"]]
Condition
  Warning:
  No event observations were detected in `truth` with event level 'Class1'.
  Warning:
  No control observations were detected in `truth` with control level '..other'.

Code
  out <- roc_auc(no_event, truth, Class1, Class2, estimator = "macro")[[
    ".estimate"]]
Condition
  Warning:
  No event observations were detected in `truth` with event level 'Class1'.
  Warning:
  No control observations were detected in `truth` with control level '..other'.

Code
  out <- roc_auc(no_event, truth, Class1, Class2, estimator = "macro_weighted")[[
    ".estimate"]]
Condition
  Warning:
  No event observations were detected in `truth` with event level 'Class1'.
  Warning:
  No control observations were detected in `truth` with control level '..other'.

Code
  out <- roc_auc(no_event, truth, Class1, Class2, estimator = "macro_weighted")[[
    ".estimate"]]
Condition
  Warning:
  No event observations were detected in `truth` with event level 'Class1'.
  Warning:
  No control observations were detected in `truth` with control level '..other'.

hand till approach throws warning and returns NaN when only 1 level has observations

Code
  out <- roc_auc_vec(x, estimate, estimator = "hand_till")
Condition
  Warning:
  x No observations were detected in `truth` for level: y.
  i Computation will proceed by ignoring those levels.

Code
  out <- roc_auc_vec(x, estimate, estimator = "hand_till")
Condition
  Warning:
  x No observations were detected in `truth` for levels: y and z.
  i Computation will proceed by ignoring those levels.

can't use case weights and hand-till method

Code
  roc_auc(hpc_cv, obs, VF:L, estimator = "hand_till", case_weights = weight)
Condition
  Error in `roc_auc()`:
  ! Can't specify both `estimator = 'hand_till'` and `case_weights`.

roc_auc() - options is deprecated

Code
  out <- roc_auc(two_class_example, truth, Class1, options = 1)
Condition
  Warning:
  The `options` argument of `roc_auc()` was deprecated in yardstick 1.0.0.
  i This argument no longer has any effect, and is being ignored. Use the pROC package directly if you need these features.

Code
  out <- roc_auc_vec(truth = two_class_example$truth, estimate = two_class_example$
    Class1, options = 1)
Condition
  Warning:
  The `options` argument of `roc_auc_vec()` was deprecated in yardstick 1.0.0.
  i This argument no longer has any effect, and is being ignored. Use the pROC package directly if you need these features.

errors with class_pred input

Code
  roc_auc_vec(cp_truth, estimate)
Condition
  Error in `roc_auc_vec()`:
  ! `truth` should not a <class_pred> object.

topepo/yardstick documentation built on April 20, 2024, 7:15 p.m.