Equality of Opportunity
Explore the concept of equality of opportunity in AI fairness, which ensures that recall rates, or true positive rates, are equal across different groups. This lesson helps you understand how to measure and interpret this metric, analyze its impact on reducing false negatives, and apply it with tools like Fairlearn to evaluate model fairness effectively.
We'll cover the following...
Equality of opportunity
Equality of opportunity can be perceived as complementary to predictive parity. While predictive parity ensures that precision for all groups is similar, equality of opportunity does the same but with recall, also called true positive rate (TPR). It is also called equal opportunity or negative error rate balance.
Let’s see the formal definition:
Where:
represents features represents model prediction based on ...