Anisotropic function estimation using multi-bandwidth Gaussian processes

Journal Article (Journal Article)

In nonparametric regression problems involving multiple predictors, there is typically interest in estimating an anisotropic multivariate regression surface in the important predictors while discarding the unimportant ones. Our focus is on defining a Bayesian procedure that leads to the minimax optimal rate of posterior contraction (up to a log factor) adapting to the unknown dimension and anisotropic smoothness of the true surface. We propose such an approach based on a Gaussian process prior with dimension-specific scalings, which are assigned carefully-chosen hyperpriors. We additionally show that using a homogenous Gaussian process with a single bandwidth leads to a sub-optimal rate in anisotropic cases.

Full Text

Duke Authors

Cited Authors

  • Bhattacharya, A; Pati, D; Dunson, D

Published Date

  • January 1, 2014

Published In

Volume / Issue

  • 42 / 1

Start / End Page

  • 352 - 381

International Standard Serial Number (ISSN)

  • 0090-5364

Digital Object Identifier (DOI)

  • 10.1214/13-AOS1192

Citation Source

  • Scopus