A0457
Title: Neyman-Pearson and equal opportunity: When efficiency meets fairness in classification
Authors: Xin Tong - University of Southern California (United States) [presenting]
Jianqing Fan - Princeton University (United States)
Shunan Yao - Hong Kong Baptist University (Hong Kong)
Yanhui Wu - University of Hong Kong (United States)
Abstract: Organizations often rely on statistical algorithms to make socially and economically impactful decisions. The fairness issues in these important automated decisions are addressed. On the other hand, economic efficiency remains instrumental in organizations survival and success. Therefore, a proper dual focus on fairness and efficiency is essential in promoting fairness in real-world data science solutions. Among the first efforts towards this dual focus, the equal opportunity (EO) constraint is incorporated into the Neyman-Pearson (NP) classification paradigm. Under this new NP-EO framework, the oracle classifier is derived, finite-sample-based classifiers are proposed that satisfy population-level fairness and efficiency constraints with high probability, and the statistical and social effectiveness of the algorithms is demonstrated on simulated and real datasets.
