In article <[EMAIL PROTECTED]>, David Jones <[EMAIL PROTECTED]> wrote:
>Maarten Speekenbrink wrote:
>> Oops, I've made a mistake in my former post. the corrected message
>is:
......................
>Johnson & Kotz suggest that others have found that a student-t
>distribution is a reasonable approximation to the logistic and you may
>find that one of these might be useful if you are looking for a
>standard type of distribution to give an idea for the shape of the
>distribution for the sum ... you would need do something like matching
>the moments.
One can get an idea about this by matching the Levy
measures, as the distributions are infinitely divisible.
>You could of course use numerical methods to do the convolution of the
>distribution functions or densities, and then plot the results.
This is RARELY a good way of doing it, especially if one
has a good means of computing the mgf. Complex inversion
of the mgf, using an appropriate contour or even steepest
descent, is a good way of getting the density or the cdf;
if the cdf is wanted, it should be done directly, not
through integrating the density.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
