Compressed Least-Squares Regression

Odalric-Ambrym Maillard 1 Rémi Munos 1
1 SEQUEL - Sequential Learning
LIFL - Laboratoire d'Informatique Fondamentale de Lille, LAGIS - Laboratoire d'Automatique, Génie Informatique et Signal, Inria Lille - Nord Europe
Abstract : We consider the problem of learning, from K data, a regression function in a linear space of high dimension N using projections onto a random subspace of lower dimension M. From any algorithm minimizing the (possibly penalized) empirical risk, we provide bounds on the excess risk of the estimate computed in the projected subspace (compressed domain) in terms of the excess risk of the estimate built in the high-dimensional space (initial domain). We show that solving the problem in the compressed domain instead of the initial domain reduces the estimation error at the price of an increased (but controlled) approximation error. We apply the analysis to Least-Squares (LS) regression and discuss the excess risk and numerical complexity of the resulting ``Compressed Least Squares Regression'' (CLSR) in terms of N, K, and M. When we choose M=O(\sqrt{K}), we show that CLSR has an estimation error of order O(\log K / \sqrt{K}).
Document type :
Conference papers
NIPS 2009, Dec 2009, Vancouver, Canada. 2009
Liste complète des métadonnées

Cited literature [26 references]  Display  Hide  Download

https://hal.inria.fr/inria-00419210
Contributor : Rémi Munos <>
Submitted on : Friday, October 30, 2009 - 2:56:20 PM
Last modification on : Thursday, January 11, 2018 - 6:22:13 AM
Document(s) archivé(s) le : Wednesday, September 22, 2010 - 1:34:57 PM

File

cls_nips.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00419210, version 2

Collections

Citation

Odalric-Ambrym Maillard, Rémi Munos. Compressed Least-Squares Regression. NIPS 2009, Dec 2009, Vancouver, Canada. 2009. 〈inria-00419210v2〉

Share

Metrics

Record views

1355

Files downloads

625