What Is Statistical Parity?

When i think about your example of a classification case, not the supervised case you're working on, it feels a little easier in some sense, to me. So then when we looked at different notions of fairness, the first and foremost that comes to our analysis was something called as statistical parity. Outcome should be independent of your membership to any group. This notion of fairness doesn't directly depend on the ground truth. And which is where we include this notion of fairness into our model. A fair ody pis a fairness ero laediction, or more complehensive understanding of what's laing on.