In article <886ii3$ser$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>Hi, a quick question:
>I am reading Silverman (Density Estimation, 1987), and hoping to apply
>it to some work I am doing.
>Let's say that I have a series of data, recurrent over several years.
>For each year, I estimate a kernel density function, and plot the
>results.
>Given the density functions for each year, is there a way to test if
>these density functions are statistically 'different?' ( I realize that
>kernel density estimation is non-parametric)
It is possible to test, but somewhat messy. The test will even
depend on the particular INFINITE-parametric estimates used; in
problems like this, while it is true that one does not have a
model with a small fixed number of parameters, it is even more
the case that there are lots of parameters, and one should approach
the problem in this manner.
I believe that when we really understand density estimation in
this manner, we will get much more efficient robust prior Bayes
procedures, which will outperform kernel estimates. Fixed width
kernels are definitely not close to optimal.
--
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, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
===========================================================================
This list is open to everyone. Occasionally, people lacking respect
for other members of the list send messages that are inappropriate
or unrelated to the list's discussion topics. Please just delete the
offensive email.
For information concerning the list, please see the following web page:
http://jse.stat.ncsu.edu/
===========================================================================
