The purpose of this paper is to investigate data-driven bandwidth sele
ction for nonparametric regression based on a double-smoothing procedu
re. It will be shown that the best convergence rate can be achieved by
kernel regression with non-negative kernels in both pilot smoothing a
nd as well as in main smoothing. The asymptotic results are given for
a naive kernel estimator with an equally spaced design, but they can a
lso be used for other kernel estimators or for locally weighted regres
sion. Three variates of data-driven bandwidth selectors for local line
ar regression are proposed. One of them, (h) over cap(DS1), is root n
consistent. The performance of these bandwidth selectors is studied th
rough simulation. They are also compared with the bandwidths selected
by the R criterion of Rice and the true ASE optimal bandwidth (h(ASE))
In spite of satisfactory performances of all bandwidth selectors, the
root n one turns out to be the best in theory as well as in practice.