yardstick (version 1.3.1)

roc_auc: Area under the receiver operator curve

Description

roc_auc() is a metric that computes the area under the ROC curve. See roc_curve() for the full curve.

Usage

roc_auc(data, ...)
# S3 method for data.frame
roc_auc(
  data,
  truth,
  ...,
  estimator = NULL,
  na_rm = TRUE,
  event_level = yardstick_event_level(),
  case_weights = NULL,
  options = list()
)
roc_auc_vec(
  truth,
  estimate,
  estimator = NULL,
  na_rm = TRUE,
  event_level = yardstick_event_level(),
  case_weights = NULL,
  options = list(),
  ...
)

Value

A tibble with columns .metric, .estimator, and .estimate and 1 row of values.

For grouped data frames, the number of rows returned will be the same as the number of groups.

For roc_auc_vec(), a single numeric value (or NA).

Arguments

data

A data.frame containing the columns specified by truth and ....

...

A set of unquoted column names or one or more dplyr selector functions to choose which variables contain the class probabilities. If truth is binary, only 1 column should be selected, and it should correspond to the value of event_level. Otherwise, there should be as many columns as factor levels of truth and the ordering of the columns should be the same as the factor levels of truth.

truth

The column identifier for the true class results (that is a factor). This should be an unquoted column name although this argument is passed by expression and supports quasiquotation (you can unquote column names). For _vec() functions, a factor vector.

estimator

One of "binary", "hand_till", "macro", or "macro_weighted" to specify the type of averaging to be done. "binary" is only relevant for the two class case. The others are general methods for calculating multiclass metrics. The default will automatically choose "binary" if truth is binary, "hand_till" if truth has >2 levels and case_weights isn't specified, or "macro" if truth has >2 levels and case_weights is specified (in which case "hand_till" isn't well-defined).

na_rm

A logical value indicating whether NA values should be stripped before the computation proceeds.

event_level

A single string. Either "first" or "second" to specify which level of truth to consider as the "event". This argument is only applicable when estimator = "binary". The default uses an internal helper that defaults to "first".

case_weights

The optional column identifier for case weights. This should be an unquoted column name that evaluates to a numeric column in data. For _vec() functions, a numeric vector, hardhat::importance_weights(), or hardhat::frequency_weights().

options

[deprecated]

No longer supported as of yardstick 1.0.0. If you pass something here it will be ignored with a warning.

Previously, these were options passed on to pROC::roc(). If you need support for this, use the pROC package directly.

estimate

If truth is binary, a numeric vector of class probabilities corresponding to the "relevant" class. Otherwise, a matrix with as many columns as factor levels of truth. It is assumed that these are in the same order as the levels of truth.

Relevant Level

There is no common convention on which factor level should automatically be considered the "event" or "positive" result when computing binary classification metrics. In yardstick, the default is to use the first level. To alter this, change the argument event_level to "second" to consider the last level of the factor the level of interest. For multiclass extensions involving one-vs-all comparisons (such as macro averaging), this option is ignored and the "one" level is always the relevant result.

Multiclass

The default multiclass method for computing roc_auc() is to use the method from Hand, Till, (2001). Unlike macro-averaging, this method is insensitive to class distributions like the binary ROC AUC case. Additionally, while other multiclass techniques will return NA if any levels in truth occur zero times in the actual data, the Hand-Till method will simply ignore those levels in the averaging calculation, with a warning.

Macro and macro-weighted averaging are still provided, even though they are not the default. In fact, macro-weighted averaging corresponds to the same definition of multiclass AUC given by Provost and Domingos (2001).

Author

Max Kuhn

Details

Generally, an ROC AUC value is between 0.5 and 1, with 1 being a perfect prediction model. If your value is between 0 and 0.5, then this implies that you have meaningful information in your model, but it is being applied incorrectly because doing the opposite of what the model predicts would result in an AUC >0.5.

Note that you can't combine estimator = "hand_till" with case_weights.

References

Hand, Till (2001). "A Simple Generalisation of the Area Under the ROC Curve for Multiple Class Classification Problems". Machine Learning. Vol 45, Iss 2, pp 171-186.

Fawcett (2005). "An introduction to ROC analysis". Pattern Recognition Letters. 27 (2006), pp 861-874.

Provost, F., Domingos, P., 2001. "Well-trained PETs: Improving probability estimation trees", CeDER Working Paper #IS-00-04, Stern School of Business, New York University, NY, NY 10012.

See Also

roc_curve() for computing the full ROC curve.

Other class probability metrics: average_precision(), brier_class(), classification_cost(), gain_capture(), mn_log_loss(), pr_auc(), roc_aunp(), roc_aunu()

Examples

# ---------------------------------------------------------------------------
# Two class example

# `truth` is a 2 level factor. The first level is `"Class1"`, which is the
# "event of interest" by default in yardstick. See the Relevant Level
# section above.
data(two_class_example)

# Binary metrics using class probabilities take a factor `truth` column,
# and a single class probability column containing the probabilities of
# the event of interest. Here, since `"Class1"` is the first level of
# `"truth"`, it is the event of interest and we pass in probabilities for it.
roc_auc(two_class_example, truth, Class1)

# ---------------------------------------------------------------------------
# Multiclass example

# `obs` is a 4 level factor. The first level is `"VF"`, which is the
# "event of interest" by default in yardstick. See the Relevant Level
# section above.
data(hpc_cv)

# You can use the col1:colN tidyselect syntax
library(dplyr)
hpc_cv %>%
  filter(Resample == "Fold01") %>%
  roc_auc(obs, VF:L)

# Change the first level of `obs` from `"VF"` to `"M"` to alter the
# event of interest. The class probability columns should be supplied
# in the same order as the levels.
hpc_cv %>%
  filter(Resample == "Fold01") %>%
  mutate(obs = relevel(obs, "M")) %>%
  roc_auc(obs, M, VF:L)

# Groups are respected
hpc_cv %>%
  group_by(Resample) %>%
  roc_auc(obs, VF:L)

# Weighted macro averaging
hpc_cv %>%
  group_by(Resample) %>%
  roc_auc(obs, VF:L, estimator = "macro_weighted")

# Vector version
# Supply a matrix of class probabilities
fold1 <- hpc_cv %>%
  filter(Resample == "Fold01")

roc_auc_vec(
   truth = fold1$obs,
   matrix(
     c(fold1$VF, fold1$F, fold1$M, fold1$L),
     ncol = 4
   )
)