scientific article; zbMATH DE number 3934219

From MaRDI portal
Publication:3706337





Let (X,Y), \((X_ i,Y_ i),i=1,2,\ldots\), be i.i.d. \(R^ p\times R^ q\)-valued random vectors with common joint distribution G(x,y) and joint density g(x,y). Let h(x) be the marginal density of X and let \(f(y| x)=g(x,y)/h(x)\) be the conditional density of Y on X. In this paper we propose the following type of conditional density estimators (called double kernel estimators) \[ f_ n(y| x)=\sum^{n}_{i=1}K_ 1(\frac{x-X_ i}{a_ n})K_ 2(\frac{y-Y_ i}{b_ n})/\{b^ q_ n\sum^{n}_{j=1}K_ 1(\frac{x-X_ j}{a_ n})\} \] where \(K_ 1\) and \(K_ 2\) are probability density functions on \(R^ p\) and \(R^ q\), respectively, and both \(a_ n\) and \(b_ n\) are sequences of small positive numbers. Denote \[ g_ n(x,y)=(na^ p_ nb^ q_ n)^{- 1}\sum^{n}_{i=1}K_ 1(\frac{x-X_ i}{a_ n})K_ 2(\frac{y-Y_ i}{b_ n}),\quad h_ n(x)=(na^ p_ n)^{-1}\sum^{n}_{i=1}K_ 1(\frac{x-X_ i}{a_ n}), \] then \(f_ n(y| x)=g_ n(x,y)/h_ n(x)\). If \(a_ n=b_ n\), \(g_ n(x,y)\) is the Rosenblatt estimator of the joint density g(x,y). In the paper we obtain both pointwise and uniform strong consistency of \(g_ n(x,y)\) and \(f_ n(y| x)\).











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 Q3706337)