adjusted_fdr {SPARRAfairness}

adjusted_fdr

Description

Estimates false discovery rate P(target=FALSE|score>cutoff,group=g) 'adjusted' for some category.

Usage

adjusted_fdr(
  scores,
  target,
  category,
  group1,
  group2,
  cutoffs = seq(min(scores, na.rm = TRUE), max(scores, na.rm = TRUE), length = 100),
  nboot = 100
)

Arguments

scores

vector of risk scores

target

vector of values of target (which risk score aims to predict)

category

vector of categories

group1

indices of group 1

group2

indices of group 2

cutoffs

score cutoffs at which to estimate metric (default 100 evenly-spaced)

nboot

number of bootstrap samples for standard error

Details

Namely, calculates

sum ( P(target=FALSE|score>cutoff,category=c,group=g)P(category=c|score<cutoff) )

where the sum is over categories c.

Value

matrix of dimension length(cutoffs)x4, with (i,2g-1)th entry the relevant fairness metric for group g at the ith cutoff value and (i,2g)th entry the approximate standard error of the (i,2g-1)th entry

Examples

# See vignette