Indicates that there is a significant discrepancy in the way different classes are being treated, which can signal bias.
Note Einstein Discovery stories are now models. We wish we could snap our fingers to update
the name everywhere, but you can expect to see the previous name in a few places until we
replace it.
Note Disparate impact is detected only for variables marked as sensitive. The alert appears as a
subtab on an insight card associated with the sensitive variable.
Actions to Consider
Investigate the variable further to determine whether it's reflecting discriminatory practices. Consider removing any biased variables from your story to remove their influence and to produce more ethical and accountable insights and models. To learn more, see Detect and Remove Bias from a Model.
Detection Methodology
Einstein Discovery raises this data alert when, for a sensitive variable, the selection data for one group is less than 80% of the group with the highest selection rate. The 80% threshold is the widely used standard for detecting disparate impact.