Re: IDL Shapiro-Wilks test ?

[Rochelle Hatton <nrh(at)imag.wsahs.nsw.gov.au>]

Karl,
Although it's probably not good to encourage your laziness,
there is some Fortran90 code at http://www.ozemail.com.au/~milleraj/
that may be converted to IDL *reasonably* painlessly, although it's pretty
long.


The IDL Astro library http://idlastro.gsfc.nasa.gov has some code
for the Kolmogorv-Smirnov test.

Rochelle

Karl Young wrote:

> Hi,
>
>     I'm playing with the idea of doing some nonnormality tests for the
> distribution of residuals for some least squares fitting problems I'm
> working on. In fumbling around trying to figure out how to do this I
> came across the name of a test called the Shapiro-Wilks test. I searched
> at LANL but didn't come up with anything and was wondering if anyone has
> written an IDL version of this test. As disgusting as this is, I'm too
> lazy to even run over to the library to find the test in a statistics
> book and code it up (the ravages of age !). I think I can hack a version
> of the Kolmogorv-Smirnov test out of the textbooks I have but I don't
> have anything that describes the Shapiro-Wilks test. Any tips greatly
> appreciated (also as I'm complete novice re. nonnormality tests, so any
> general pointers on these welcome as well),
>
> -- KY
>
> Karl Young
> Research Physicist
> UC San Francisco
> kyoung@itsa.ucsf.edu

--
Computers are useless. They can only give you answers - Pablo Picasso

Rochelle Hatton
Department of Nuclear Medicine and Ultrasound
Westmead Hospital, Sydney