Skip to contents

recall estimates the recall (a.k.a. sensitivity, true positive rate -TPR-, or hit rate) for a nominal/categorical predicted-observed dataset.

TPR alternative to recall().

sensitivity alternative to recall().

hitrate alternative to recall().

FNR estimates false negative rate (or false alarm, or fall-out) for a nominal/categorical predicted-observed dataset.

Usage

recall(
  data = NULL,
  obs,
  pred,
  atom = FALSE,
  pos_level = 2,
  tidy = FALSE,
  na.rm = TRUE
)

TPR(
  data = NULL,
  obs,
  pred,
  atom = FALSE,
  pos_level = 2,
  tidy = FALSE,
  na.rm = TRUE
)

sensitivity(
  data = NULL,
  obs,
  pred,
  atom = FALSE,
  pos_level = 2,
  tidy = FALSE,
  na.rm = TRUE
)

hitrate(
  data = NULL,
  obs,
  pred,
  atom = FALSE,
  pos_level = 2,
  tidy = FALSE,
  na.rm = TRUE
)

FNR(
  data = NULL,
  obs,
  pred,
  atom = FALSE,
  pos_level = 2,
  tidy = FALSE,
  na.rm = TRUE
)

Arguments

data

(Optional) argument to call an existing data frame containing the data.

obs

Vector with observed values (character | factor).

pred

Vector with predicted values (character | factor).

atom

Logical operator (TRUE/FALSE) to decide if the estimate is made for each class (atom = TRUE) or at a global level (atom = FALSE); Default : FALSE. When dataset is "binomial" atom does not apply.

pos_level

Integer, for binary cases, indicating the order (1|2) of the level corresponding to the positive. Generally, the positive level is the second (2) since following an alpha-numeric order, the most common pairs are (Negative | Positive), (0 | 1), (FALSE | TRUE). Default : 2.

tidy

Logical operator (TRUE/FALSE) to decide the type of return. TRUE returns a data.frame, FALSE returns a list; Default : FALSE.

na.rm

Logic argument to remove rows with missing values (NA). Default is na.rm = TRUE.

Value

an object of class numeric within a list (if tidy = FALSE) or within a data frame (if tidy = TRUE).

Details

The recall (a.k.a. sensitivity or true positive rate -TPR-) is a non-normalized coefficient that represents the ratio between the correctly predicted cases (true positives -TP-) to the total number of actual observations that belong to a given class (actual positives -P-).

For binomial cases, \(recall = \frac{TP}{P} = \frac{TP}{TP + FN} \)

The recall metric is bounded between 0 and 1. The closer to 1 the better. Values towards zero indicate low performance. It can be either estimated for each particular class or at a global level.

Metrica offers 4 identical alternative functions that do the same job: i) recall, ii) sensitivity, iii) TPR, and iv) hitrate. However, consider when using metrics_summary, only the recall alternative is used.

The false negative rate (or false alarm, or fall-out) is the complement of the recall, representing the ratio between the number of false negatives (FN) to the actual number of positives (P). The FNR formula is:

\(FNR = 1 - recall = 1 - TPR = \frac{FN}{P}\)

The fpr is bounded between 0 and 1. The closer to 0 the better. Low performance is indicated with fpr > 0.5.

For the formula and more details, see online-documentation

References

Ting K.M. (2017) Precision and Recall. In: Sammut C., Webb G.I. (eds) Encyclopedia of Machine Learning and Data Mining. Springer, Boston, MA. doi:10.1007/978-1-4899-7687-1_659

Ting K.M. (2017). Sensitivity. In: Sammut C., Webb G.I. (eds) Encyclopedia of Machine Learning and Data Mining. Springer, Boston, MA. doi:10.1007/978-1-4899-7687-1_751

Trevethan, R. (2017). Sensitivity, Specificity, and Predictive Values: Foundations, Pliabilities, and Pitfalls _ in Research and Practice. Front. Public Health 5:307_ doi:10.3389/fpubh.2017.00307

Examples

# \donttest{
set.seed(123)
# Two-class
binomial_case <- data.frame(labels = sample(c("True","False"), 100, 
replace = TRUE), predictions = sample(c("True","False"), 100, replace = TRUE))

# Multi-class
multinomial_case <- data.frame(labels = sample(c("Red","Blue", "Green"), 100, 
replace = TRUE), predictions = sample(c("Red","Blue", "Green"), 100, replace = TRUE))

# Get recall estimate for two-class case at global level
recall(data = binomial_case, obs = labels, pred = predictions, tidy = TRUE)
#>      recall
#> 1 0.4561404

# Get FNR estimate for two-class case at global level
FNR(data = binomial_case, obs = labels, pred = predictions, tidy = TRUE)
#>         FNR
#> 1 0.5438596

# Get recall estimate for each class for the multi-class case at global level
recall(data = multinomial_case, obs = labels, pred = predictions, tidy = TRUE, 
atom = FALSE)
#>      recall
#> 1 0.2770092

# Get recall estimate for the multi-class case at a class-level
recall(data = multinomial_case, obs = labels, pred = predictions, tidy = TRUE,
atom = TRUE)
#>          recall
#> Blue  0.2727273
#> Green 0.2173913
#> Red   0.3409091
# }