[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
kolmogorov smirnov test. Normal distribution
From: |
John W. Eaton |
Subject: |
kolmogorov smirnov test. Normal distribution |
Date: |
Thu, 24 Apr 2008 13:34:09 -0400 |
On 24-Apr-2008, A. Kalten wrote:
|
| >> [pval, ks] = kolmogorov_smirnov_test(X,'normal',mean(X),var(X), "<>");
| >> pval
| >> pval = 0.92970
| >> ks
| >> ks = 0.54297
|
| > In minitab (see above) I have got KS=0.140 which is very different from
| > octave value for this.
| > And I still can't get why. Do I have a crappy statistical software
| > (minitab)? Or do I use octave incorrectly?
| > Or what?
|
| The octave source code for the KS test shows that the algorithm
| used differs from the algorithm that may be used by other statistics
| packages, such as R and presumably minitab. In particular, the ks
| statistic is scaled by sqrt(n), where n is the number of data points.
| Also, the octave code uses a slightly different method to determine
| the maximum difference between the cumulative distributions.
|
| The reasons for these differing algorithms are somewhat complex
| mathematically, but if you are interested there is a lot of
| information available on the web.
|
| It would be nice if each octave function had references included
| within the source file so that the theoretical background of the
| function could be researched. It's never a good idea to use any
| function in a blind manner.
If you look at the source for the k-s function in Octave, you'll see
that it was written by someone who is now a member of the R-core
development group. Please remember that nearly everyone working on
Octave is a volunteer. It's not that we are trying to do a bad job
here. If you know better, then perhaps you could contribute a fix?
jwe