Journal Article

Efficient Gaussian process regression for large datasets

Anjishnu Banerjee, David B. Dunson and Surya T. Tokdar

in Biometrika

Published on behalf of Biometrika Trust

Volume 100, issue 1, pages 75-89
Published in print March 2013 | ISSN: 0006-3444
Published online December 2012 | e-ISSN: 1464-3510 | DOI: http://dx.doi.org/10.1093/biomet/ass068
Efficient Gaussian process regression for large datasets

More Like This

Show all results sharing this subject:

  • Probability and Statistics

GO

Show Summary Details

Preview

Gaussian processes are widely used in nonparametric regression, classification and spatiotemporal modelling, facilitated in part by a rich literature on their theoretical properties. However, one of their practical limitations is expensive computation, typically on the order of n3 where n is the number of data points, in performing the necessary matrix inversions. For large datasets, storage and processing also lead to computational bottlenecks, and numerical stability of the estimates and predicted values degrades with increasing n. Various methods have been proposed to address these problems, including predictive processes in spatial data analysis and the subset-of-regressors technique in machine learning. The idea underlying these approaches is to use a subset of the data, but this raises questions concerning sensitivity to the choice of subset and limitations in estimating fine-scale structure in regions that are not well covered by the subset. Motivated by the literature on compressive sensing, we propose an alternative approach that involves linear projection of all the data points onto a lower-dimensional subspace. We demonstrate the superiority of this approach from a theoretical perspective and through simulated and real data examples.

Keywords: Bayesian regression; Compressive sensing; Dimensionality reduction; Gaussian process; Random projection

Journal Article.  0 words. 

Subjects: Probability and Statistics

Full text: subscription required

How to subscribe Recommend to my Librarian

Users without a subscription are not able to see the full content. Please, subscribe or login to access all content.