FEERCI is an opinionated, easy-to-use package for calculating EERs and non-parametric confidence intervals efficiently. It offers a single method, feerci.feerci, that returns both an EER and CI for provided impostor and genuine scores. The only dependency is numpy.

pip install feerci

• Switched output arguments around, to make more intuitive sense

• Initial release of package

FEERCI is distributed under the MIT license

0.2.0

Calculating just an EER:

import feerci
import numpy as np
impostors = np.random.rand(100)
genuines = np.random.rand(100)
eer,_,_,_ = feerci.feerci(impostors,genuines,is_sorted=False,m=-1)

Calculating an EER and 95% confidence interval (the default) on 10000 bootstrap iterations (the default):

eer,ci_lower,ci_upper,bootstrapped_eers = feerci.feerci(impostors,genuines,is_sorted=False)

Calculating an EER and 99% confidence interval on 10000 bootstrap iterations (the default):

eer,ci_lower,ci_upper,bootstrapped_eers = feerci.feerci(impostors,genuines,is_sorted=False,ci=.99)

Calculating an EER and 99% confidence interval on 1000 bootstrap iterations:

eer,ci_lower,ci_upper,bootstrapped_eers = feerci.feerci(impostors,genuines,is_sorted=False,m=1000,ci=.99)