Speed of convergence in nonparametric kernel estimation of a regression function and its derivatives (Q1060511)

The objective in nonparametric regression is to infer a function g(x) and its pth order derivatives \(g^{(p)}(x)\), \(p\geq 1\) fixed, on the basis of a finite collection of pairs \(\{x_ i,g(x_ i)+Z_ i\}^ n_{i=1}\), where the noise components \(Z_ i\) satisfy certain modest assumptions and the domain points \(x_ i\) are selected non-randomly. This paper exhibits a new class of kernel estimates \(g_ n^{(p)}\), \(p\geq 0\) fixed. The main theoretical results of this study are the rates of convergence obtained for mean square and strong consistency of \(g_ n^{(p)}\), each of them being uniform on the (0,1).