"David Jones" <[EMAIL PROTECTED]> wrote in message news:<[EMAIL PROTECTED]>...
> The above has overlooked a number of possibilities:
> 
> (i) use of a Kernel estimate for the distribution function (as a
> non-parametric estimate alternative to just using the sample
> quantile).

I once asked Matt Wand (of Wand, MP & MC Jones, Kernel Smoothing, 
Chapman & Hall, London, 1995) about this and he suggested that
generally you don't want to do this and would usually be better
just to use the sample quantile (if you can specify enough about
your distribution that you can be confident your kernel estimate
of the quantile is better, you can often use that information
more directly to get a better directly-based-on-sample estimate).

(That's not to say you'd never use a kernel based estimate.)

Glen
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to