Conditional regression for single-index models

DOI10.3150/22-BEJ1482MaRDI QIDQ6335420arXivFDO

Authors Alessandro Lanteri, Mauro Maggioni, Stefano Vigogna

Publication date 23 February 2020

Published in arXiv

Full work available at URL https://doi.org/10.3150/22-BEJ1482

Mathematics Subject Classification ID

Nonparametric inference (62Gxx) Multivariate analysis (62Hxx) Linear inference, regression (62Jxx)

arXiv classification

math.STstat.TH

Abstract: The single-index model is a statistical model for intrinsic regression where responses are assumed to depend on a single yet unknown linear combination of the predictors, allowing to express the regression function as $mathbbE[Y|X]=f(langlev,Xangle)$ for some unknown emph{index} vector $v$ and emph{link} function $f$. Conditional methods provide a simple and effective approach to estimate $v$ by averaging moments of $X$ conditioned on $Y$, but depend on parameters whose optimal choice is unknown and do not provide generalization bounds on $f$. In this paper we propose a new conditional method converging at $sqrtn$ rate under an explicit parameter characterization. Moreover, we prove that polynomial partitioning estimates achieve the $1$-dimensional min-max rate for regression of H"older functions when combined to any $sqrtn$-convergent index estimator. Overall this yields an estimator for dimension reduction and regression of single-index models that attains statistical optimality in quasilinear time.

This page was built for publication: Conditional regression for single-index models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6335420)