Uniformity in \(h\) in the functional limit law for the increments of the empirical process indexed by functions (Q1776845)

The article studies the almost sure convergence of kernel density estimators uniformly in the bandwidth \(h_n \leq h \leq \tilde h_n\) and the kernel \(K\), where \(h_n\) and \(\tilde h_n\) satisfy certain growth conditions and \(K \in {\mathcal F}\), with \({\mathcal F}\) denoting a family of functions fulfilling appropriate boundedness, support and entropy assumptions.