Abstract
It is well known that data-driven regression smoothing parameters ħ based on cross-validation and related methods exhibit a slow rate of convergence to their optimum. In an earlier article we showed that this rate can be as slow as n –1/10; that is, for a bandwidth ħ 0 optimizing the averaged squared error, n 1/10 (ħ — ħ 0)/ħ 0 tends to an asymptotic normal distribution. In this article we consider mean averaged squared error optimal bandwidths h 0. This (nonrandom) smoothing parameter can be approximated much faster. We use the technique of double smoothing to show that there is an ħ such that, under certain conditions, n 1/2(ħ − h 0)/h 0 tends to an asymptotic normal distribution.