Distribution function estimation: Adaptive smoothing (Q1366480)

In this paper, a sequel to our paper Math. Methods Stat. 5, No. 1, 1-31 (1996; Zbl 0853.62028), we continue to study the problem of the second order minimax asymptotic estimation of an unknown distribution function (d.f.). Our attention here is restricted to Sobolev functional classes. Let \(X^{(n)}= (X_1,\dots, X_n)\) be independent random observations with a common d.f. \(F(x)\), \(x\in \mathbb{R}^1\). It is assumed that \(F(x)\) is unknown. Our goal is to estimate \(F(x)\) based on the observations \(X^{(n)}\) in such a way as to minimize the integrated mean square error.