scientific article; zbMATH DE number 17214

From MaRDI portal
Publication:3973911





[For the entire collection see Zbl 0722.00032.] Consider an i.i.d. sample \((X_ i,Y_ i)^ n_{i=1}\) from an \({\mathbb{R}}^ p\times{\mathbb{R}}\) valued random vector, satisfying \(Y_ i=X_ i\theta+\varepsilon_ i\sigma(X_ i)\), \(1\leq i\leq n\), where \(\theta\) is a \(p\)-dimensional parameter, \(\sigma(.)\) an unknown function and the \(\varepsilon_ i\) are, conditionally on \(X_ i\), symmetric and have variance 1. A set of regularity conditions is given under which \(M\)- estimators for \(\theta\) based on certain nonparametric estimates of \(\sigma(.)\) are asymptotically equivalent to those using the true variance function. A simulation study gives some information about the finite sample performance.











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 Q3973911)