Bayesian model uncertainty and selection under differential privacy

V. Peña

In this talk, we present differentially private methods for model averaging and selection for normal linear models. The methods are based on mixtures of g-priors. The procedures are asymptotically consistent and straightforward to implement with existing software.

Keywords: Variable selection Differential Privacy Bayesian Methods

D. Rossell, P. Rognon, P. Zwiernick

G. Garcia-Donato, M. E. Castellanos Nueda

S. Cabras, M. E. Castellanos Nueda, A. Forte Deltell, G. Garcia-Donato, A. Quirós Carretero