Table 1. Mathematical Definitions of 3 Common Fairness Metrics.

Fairness metric	Definitiona	Explanation with sample of use in Alzheimer disease
Equal opportunity	True positive rates are the same across groups: P(Ŷ = 1|A = 0, Y = 1) = P(Ŷ = 1|A = 1, Y = 1)	The probability of correctly predicting that an individual progresses to AD is the same for groups defined by a protected attribute, such as race
Equal odds	True positive rates and false positive rates are the same across groups: P(Ŷ = 1|A = 0, Y = y) = P(Ŷ = 1|A = 1, Y = y), y ∈ {0,1}	The probability of correctly predicting that an individual progresses to AD and the probability of incorrectly predicting progression to AD for those who do not are the same for groups defined by a protected attribute, such as race
Demographic parity	Equal probability of being classified with the positive label: P(Ŷ = 1|A = 0) = P(Ŷ = 1|A = 1)	The proportion of individuals predicted to progress to AD is the same across groups defined by a protected attribute, such as race

^a Y indicates the observed outcome, Ŷ a prediction of Y, and A a binary protected attribute.