auprc: Calculate area under precision recall curve (AUPRC)

In jweile/yogiroc: Draw ROC and PRC curves

Calculate area under precision recall curve (AUPRC)

Description

Balancing concept by Yingzhou Wu and Fritz Roth (Wu et al, unpublished)

Usage

auprc(yr2, monotonized = TRUE, balanced = FALSE)

Arguments

yr2	the yogiroc2 object
monotonized	whether or not use a monotonized PRC curve
balanced	whether or not to use prior-balancing

Value

a numerical vector with the AUPRC values for each predictor

Examples

#generate fake data
truth <- c(rep(TRUE,10),rep(FALSE,8))
scores <- cbind(
  pred1=c(rnorm(10,1,0.2),rnorm(8,.9,0.1)),
  pred2=c(rnorm(10,1.1,0.2),rnorm(8,.9,0.2))
)
#create yogiroc2 object
yrobj <- yr2(truth,scores)
#calculate AUPRC
auprc(yrobj)
#calculate non-monotonized AUPRC
auprc(yrobj,monotonized=FALSE)
#calculate balanced AUPRC
auprc(yrobj,balanced=TRUE)

jweile/yogiroc documentation built on Jan. 15, 2024, 2:47 a.m.