scientific article; zbMATH DE number 1762677

From MaRDI portal
Publication:4538828





In nonparametric regression models we observe \(y_i=g(\xi_i)+\epsilon_i,\;i=1,\dots,n,\) where \(\epsilon_i\) are i.i.d. r.v.s with zero mean, finite variance and density \(p_{\epsilon}\). The design points or variables \(\xi_1,\dots,\xi_n\) are known. In nonparametric errors-in-variables regression models the design points or variables \(\xi_1,\dots,\xi_n\) are unknown. It is supposed that the equation \(x_i=\xi_i+\delta_i\) holds, where \(\delta_i\) are i.i.d. r.v.s with zero mean and density \(p_{\delta}\). A model is called structural errors-in-variables model if the \(\xi_i\) are i.i.d. r.v.s, independent on \(\xi_j\) and \(\delta_j\), with density \(p_{\xi}\). The problem is to estimate the unknown regression function \(g\) for errors-in-variables models from observations of the design points \(x_1,\dots,x_n\). The author proposes a general procedure for the adjustment of nonparametric regression estimators to models with errors in variables.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4538828)