Skip to main content
. 2020 Dec 17;3:559927. doi: 10.3389/frai.2020.559927

FIGURE 4.

FIGURE 4

Q-Finder works in 4 main stages: an exhaustive generation of candidate subgroups, a ranking of candidate subgroups via an evaluation of their empirical credibility, a selection of the best candidates (taking into account the redundancy between subgroups), and then an assessment of subgroups’ credibility on one or more test datasets.