getprc {SPARRAfairness}R Documentation

getprc()

Description

Comprehensive plotting function for precision-recall curve. Also calculates AUPRC and standard error.

Usage

getprc(y, ypred, cv = NULL, res = 100)

Arguments

y

class labels, 0/1 or logical

ypred

predictions Pr(Y=1), numeric vector

cv

cross-validation fold assignments, if relevant. Changes estimate of standard error.

res

resolution. Returns this many equally-spaced points along the curve. Set res to null to return all points.

Details

Rather than returning points corresponding to every cutoff, only returns a representative sample of equally-spaced points along the curve.

Does not plot anything. Object can be plotted in a default way.

Value

list containing: ppv, ppv for res points in every cv fold; sens, sensitivity for res points in every cv fold; auc, areas under the curve for each fold and average (note length is 1 greater than number of CV folds); se, standard error for AUC in each fold and standard error for average auc (note length is 1 greater than number of CV folds)

Examples

# See vignette

[Package SPARRAfairness version 0.0.0.1 Index]