specificity estimates the specificity (a.k.a. selectivity,
or true negative rate -TNR-)
for a nominal/categorical predicted-observed dataset.
selectivity alternative to specificity().
TNR alternative to specificity().
FPR estimates the false positive rate (a.k.a fall-out or false alarm)
for a nominal/categorical predicted-observed dataset.
Usage
specificity(
data = NULL,
obs,
pred,
atom = FALSE,
pos_level = 2,
tidy = FALSE,
na.rm = TRUE
)
selectivity(
data = NULL,
obs,
pred,
atom = FALSE,
pos_level = 2,
tidy = FALSE,
na.rm = TRUE
)
TNR(
data = NULL,
obs,
pred,
atom = FALSE,
pos_level = 2,
tidy = FALSE,
na.rm = TRUE
)
FPR(
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 specificity (or selectivity, or true negative rate-TNR-) is a non-normalized coefficient that represents the ratio between the correctly negative predicted cases (or true negative -TN- for binary cases) to the total number of actual observations not belonging to a given class (actual negatives -N- for binary cases).
For binomial cases, \(specificity = \frac{TN}{N} = \frac{TN}{TN+FP}\)
The specificity metric is bounded between 0 and 1. The closer to 1 the better.
Values towards zero indicate low performance. For multinomial cases, it can be
either estimated for each particular class or at a global level.
Metrica offers 3 identical alternative functions that do the same job: i) specificity,
ii) selectivity, and iii) TNR. However, consider
when using metrics_summary, only the specificity alternative is used.
The false positive rate (or false alarm, or fall-out) is the complement of the
specificity, representing the ratio between the number of false positives (FP)
to the actual number of negatives (N). The FPR formula is:
\(FPR = 1 - specificity = 1 - TNR = \frac{FP}{N}\)
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) Sensitivity and Specificity. In: Sammut C., Webb G.I. (eds) Encyclopedia of Machine Learning and Data Mining. Springer, Boston, MA. doi:10.1007/978-0-387-30164-8_752
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 specificity and FPR estimates for two-class case
specificity(data = binomial_case, obs = labels, pred = predictions, tidy = TRUE)
#> spec
#> 1 0.5348837
FPR(data = binomial_case, obs = labels, pred = predictions, tidy = TRUE)
#> FPR
#> 1 0.4651163
# Get specificity estimate for each class for the multi-class case
specificity(data = multinomial_case, obs = labels, pred = predictions, tidy = TRUE)
#> spec
#> 1 0.639465
# Get specificity estimate for the multi-class case at a global level
specificity(data = multinomial_case, obs = labels, pred = predictions, tidy = TRUE)
#> spec
#> 1 0.639465
# }