Kernel dimension reduction in regression
arXiv:0908.1854 · doi:10.1214/08-AOS637
Abstract
We present a new methodology for sufficient dimension reduction (SDR). Our methodology derives directly from the formulation of SDR in terms of the conditional independence of the covariate from the response , given the projection of on the central subspace [cf. J. Amer. Statist. Assoc. 86 (1991) 316--342 and Regression Graphics (1998) Wiley]. We show that this conditional independence assertion can be characterized in terms of conditional covariance operators on reproducing kernel Hilbert spaces and we show how this characterization leads to an -estimator for the central subspace. The resulting estimator is shown to be consistent under weak conditions; in particular, we do not have to impose linearity or ellipticity conditions of the kinds that are generally invoked for SDR methods. We also present empirical results showing that the new methodology is competitive in practice.
Published in at http://dx.doi.org/10.1214/08-AOS637 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)
Cited by in corpus (6)
- A general theory for nonlinear sufficient dimension reduction: Formulation and estimation
- Cross-Domain Object Matching with Model Selection
- Discussion of: Brownian distance covariance
- Comment on "Support Vector Machines with Applications"
- Sufficient Component Analysis for Supervised Dimension Reduction
- Learning sparse gradients for variable selection and dimension reduction