RT journal article
T1 Empirical Bayes fairness in linear regression
A1 Carrizosa, Emilio
A1 Jiménez-Llamas, Rafael
A1 Ramírez Cobo, Josefa
A2 Estadística e Investigación Operativa
K1 fairness
K1 machine learning
K1 linear regression model
K1 Normal-Gamma prior,
K1 empirical Bayes
K1 unfairness measure
K1 Gram-Schmidt process
AB Bias in data may lead to prediction procedures which discriminate individuals from sensitive groups. In this paper we propose a Bayesian method for parameter estimation in the linear regression model taking fairness into ac- count. For a given prior structure, namely, the Normal-Gamma prior, for a specific choice of unfairness measure (which does not require the use of private -sensitive- information), our method computes a constrained empirical Bayes estimator that induces fairness in posterior predictions. Specifically, the model hyper-parameters are calculated as the optimal solution of a constrained optimization problem: that in which the marginal likelihood is maximized in a region where the unfairness measure is upper bounded and thus, controlled. This guarantees that average posterior predictions are, at most, as unfair as the upper bound establishes. Experi- ments with synthetic and real data sets show the competitiveness of our approach in terms of fairness in the obtained solution and prediction error.
YR 2024
FD 2024
LK http://hdl.handle.net/10498/32822
UL http://hdl.handle.net/10498/32822
LA eng
DS Repositorio Institucional de la Universidad de Cádiz
RD 08-nov-2026