Rich Strauss wrote:
> In my fields of interest (ecology and evolutionary biology), it is
becoming
> increasing common to refer to two "kinds" of bootstrapping: nonparametric
> bootstrapping, in which replicate samples are drawn randomly with
> replacement from the original sample; and parametric bootstrapping, in
> which samples are drawn randomly from a (usually normal) distribution
> having the same mean and variance as the original sample.
I suppose the justification for the latter is that it avoids certain
dissimilarities between the true and empirical distributions (eg,
granularity). Presumably smoothing the empirical distribution would have
the same effect, with less violence to the true shape of the distribution.
A kernel smoother would be particularly easy, as it would correspond to
adding a small random perturbation to each element of the bootstrap sample.
Does anybody know a good source on this? In particular, how does one decide
on the shape and size of the perturbation? Or are there good reasons not to
do this?
-Robert Dawson
