Skip to main content
. 2020 May 12;3:33. doi: 10.3389/frai.2020.00033

Figure 4.

Figure 4

Accuracy and fairness (with respect to equality of opportunity) for various methods on ProPublica dataset. (A): using race as the sensitive attribute; (B): using gender. A completely fair model would achieve a value of 1.0 in the x-axis. See Figures 5A,B on how these choices of PR setting translate to TPRs = 0 vs TPRs = 1.