How to calculate 95% CI for area under the precision recall curve in R

Question

I calculate area under the precision recall curve using the Mleval package. The Mleval package provides 95% CI for the AUROC but not AUPRC. Is there a package that provides 95% CI for AUPRC?

Answer 1

The precrec library gives an AUC for precision-recall as well as ROC, as shown in this example with dummy data:

library(precrec)

set.seed(1)

## Create sample datasets with 100 positives and 100 negatives
samps <- create_sim_samples(2, 100, 100, "good_er")

## Single model & multiple test datasets
smmdat <- mmdata(samps[["scores"]], samps[["labels"]],
                 modnames = samps[["modnames"]],
                 dsids = samps[["dsids"]])

mod <- evalmod(smmdat)
auc_ci(mod)
#>   modnames curvetypes      mean        error lower_bound upper_bound n
#> 1  good_er        ROC 0.8192000 0.0007839856   0.8184160    0.819984 2
#> 2  good_er        PRC 0.8603505 0.0154645417   0.8448859    0.875815 2

^{Created on 2022-03-06 by the reprex package (v2.0.1)}