Skip to contents

Concordance indexes

Source: R/concordace.R
concordance_indexes.Rd

Concordance derived indexes allow calculation and explanation of area under ROC curve in a specific region. They use a dual perspective since they consider both TPR and FPR ranges which enclose the region of interest.

cp_auc() applies concordan partial area under curve (CpAUC), while ncp_auc() applies its normalized version by dividing by the total area.

Usage

cp_auc(
  data = NULL,
  response,
  predictor,
  lower_threshold,
  upper_threshold,
  ratio,
  .condition = NULL
)

ncp_auc(
  data = NULL,
  response,
  predictor,
  lower_threshold,
  upper_threshold,
  ratio,
  .condition = NULL
)

Arguments

data

A data.frame or extension (e.g. a tibble) containing values for predictors and response variables.

response

A data variable which must be a factor, integer or character vector representing the prediction outcome on each observation (Gold Standard).

If the variable presents more than two possible outcomes, classes or categories:

  • The outcome of interest (the one to be predicted) will remain distinct.

  • All other categories will be combined into a single category.

New combined category represents the "absence" of the condition to predict. See .condition for more information.

predictor

A data variable which must be numeric, representing values of a classifier or predictor for each observation.

lower_threshold, upper_threshold

Two numbers between 0 and 1, inclusive. These numbers represent lower and upper bounds of the region where to apply calculations.

ratio

Ratio or axis where to apply calculations.

  • If "tpr", only points within the specified region of TPR, y axis, will be considered for calculations.

  • If "fpr", only points within the specified region of FPR, x axis, will be considered for calculations.

.condition

A value from response that represents class, category or condition of interest which wants to be predicted.

If NULL, condition of interest will be selected automatically depending on response type.

Once the class of interest is selected, rest of them will be collapsed in a common category, representing the "absence" of the condition to be predicted.

See vignette("selecting-condition") for further information on how automatic selection is performed and details on selecting the condition of interest.

Value

A numeric value representing index score for the partial area under ROC curve.

References

Carrington, André M., et al. A new concordant partial AUC and partial c statistic for imbalanced data in the evaluation of machine learning algorithms. BMC medical informatics and decision making 20 (2020): 1-12.

Examples

# Calculate cp_auc of Sepal.Width as a classifier of setosa especies in
# FPR = (0, 0.1)
cp_auc(
  iris,
  response = Species,
  predictor = Sepal.Width,
  lower_threshold = 0,
  upper_threshold = 0.1,
  ratio = "fpr"
)
#> ℹ Lower 0 and upper 0.1 thresholds already included in points
#> • Skipping lower and upper threshold interpolation
#> [1] 0.3446
# Calculate ncp_auc of Sepal.Width as a classifier of setosa especies in
# FPR = (0, 0.1)
ncp_auc(
  iris,
  response = Species,
  predictor = Sepal.Width,
  lower_threshold = 0,
  upper_threshold = 0.1,
  ratio = "fpr"
)
#> ℹ Lower 0 and upper 0.1 thresholds already included in points
#> • Skipping lower and upper threshold interpolation
#> ℹ Lower 0 and upper 0.1 thresholds already included in points
#> • Skipping lower and upper threshold interpolation
#> [1] 0.9068421