Many binary classification metrics {Rfast2}

Many binary classification metrics

Description

Many binary classification metrics.

Usage

colaccs(group, preds)
colsens(group, preds)
colspecs(group, preds)
colprecs(group, preds)
colfscores(group, preds)
colfbscores(group, preds, b)
colfmis(group, preds)

Arguments

group

A numerical vector with two values, 0 and 1.

preds

A numerical matrix with scores, probabilities or any other measure.

b

The \beta parameter in the F_{\beta}-score.

Details

The accuracies, sensitivities, specificities, precisions, F-scores, F_{\beta}-scores and the Fowlkes-Mallows index are calculated column-wise. The colaccs is the only metric that can be used with a multinomial response as well.

Value

A vector with length equal to the number of columns of the "preds" argument containing the relevant values computed for each column.

Author(s)

Michail Tsagris.

R implementation and documentation: Michail Tsagris [e-mail].

References

https://en.wikipedia.org/wiki/Sensitivity_and_specificity

https://en.wikipedia.org/wiki/Precision_and_recall

See Also

colmses, bernoulli.nb, bigknn.cv

Examples

## 20 variables, hence 20 accuracies will be calculated
ina <- rbinom(100, 1, 0.6)
x <- matrix( rnorm(100 * 20), ncol = 20 )
a <- colaccs(ina, x)