specificity: Specificity

In brandon-mosqueda/SKM: Sparse Kernels Methods

Specificity

Description

Given the observed and predicted values of categorical data (of any number of classes) computes the specificity, the metric that evaluates a model's ability to predict true negatives of each available category.

Usage

specificity(observed, predicted, positive_class = NULL, remove_na = TRUE)

Arguments

observed	(factor) The observed values. It has to have the same length as predicted.
predicted	(factor) The observed values. It has to have the same length as observed.
positive_class	(character(1)) The name of the class (level) to be taken as reference as the positive class. This parameter is only used for binary variables. NULL by default which uses the second class in the union of the classes (levels) in observed and predicted.
remove_na	(logical(1)) Should NA values be removed?. TRUE by default.

Details

Given the following binary confusion matrix:

Specificity is computed as:

Value

For binary data a single value is returned, for more than 2 categories a vector of sensitivities is returned, one per each category.

See Also

Other categorical_metrics: accuracy(), brier_score(), categorical_summary(), confusion_matrix(), f1_score(), kappa_coeff(), math_mode(), matthews_coeff(), pccc(), pcic(), pr_auc(), precision(), recall(), roc_auc(), sensitivity()

Examples

## Not run: 
specificity(factor(c("a", "b")), factor(c("a", "b")))
specificity(factor(c("a", "b")), factor(c("b", "a")))
specificity(factor(c("a", "b")), factor(c("b", "b")))
specificity(factor(c(TRUE, FALSE)), factor(c(FALSE, TRUE)))
specificity(factor(c("a", "b", "a")), factor(c("b", "a", "c")))

## End(Not run)

brandon-mosqueda/SKM documentation built on Feb. 8, 2025, 5:24 p.m.