Linear regression estimation in non-linear single index models

arXiv1612.07704MaRDI QIDQ6281168FDOQ6281168

Authors: Fadoua Balabdaoui, Gian-Andrea Thanei 

Publication date: 22 December 2016

Abstract: In this article, we consider the problem of estimating the index parameter $alph{a}_0$ in the single index model $E[Y|X]=f_0(alph{a}_0^{T}X)$ with $f_0$ the unknown ridge function defined on $mathbbR$, $X$ a d-dimensional covariate and $Y$ the response. We show that when $X$ is Gaussian, then $alph{a}_0$ can be consistently estimated by regressing the observed responses $Y_i$, $i=1,\mathrm{...},n$ on the covariates $X_1,\mathrm{...},X_n$ after centering and rescaling. The method works without any additional smoothness assumptions on $f_0$ and only requires that $cov(f_0(alph{a}_0^{T}X),alph{a}_0^{T}X)eq0$, which is always satisfied by monotone and non-constant functions $f_0$. We show that our estimator is asymptotically normal and give the expression with its asymptotic variance. The approach is illustrated through a simulation study.

Mathematics Subject Classification ID

General nonlinear regression (62J02)

This page was built for publication: Linear regression estimation in non-linear single index models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6281168)