The Annals of Statistics

An Empirical Bayes Approach to Multiple Linear Regression

Serge L. Wind

Full-text: Open access


We consider estimation (subject to quadratic loss) of the vector of coefficients of a multiple linear regression model in which the error vector is assumed to have 0 mean and covariance matrix $\sigma^2 I$ but is not assumed to take on a specific parametric form, e.g., Normal. The vector of coefficients is taken to be randomly distributed according to some unknown prior. Restricted minimax solutions are exhibited relative to equivalence classes on the space of all prior probability distributions which group distributions with the same specified moments. In the context of the classic Empirical Bayes formulation, we determine restricted asymptotically optimal estimators--i.e., decision functions whose Bayes risks converge to the risk of the restricted minimax decision at each component stage.

Article information

Ann. Statist., Volume 1, Number 1 (1973), 93-103.

First available in Project Euclid: 25 October 2007

Permanent link to this document

Digital Object Identifier

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

Primary: 62C10: Bayesian problems; characterization of Bayes procedures
Secondary: 62J05: Linear regression 62G05: Estimation

Empirical Bayes regression Stein-James estimators restricted minimax


Wind, Serge L. An Empirical Bayes Approach to Multiple Linear Regression. Ann. Statist. 1 (1973), no. 1, 93--103. doi:10.1214/aos/1193342385.

Export citation