From octave-sources-request at bevo dot che dot wisc dot edu Thu Aug 30 20:51:13 2001
Subject: Re: Kolmogorov-Smirnov CDF
From: "James R. Van Zandt" <jrv at vanzandt dot mv dot com>
To: Kurt dot Hornik at ci dot tuwien dot ac dot at
Cc: octave-sources at bevo dot che dot wisc dot edu
Date: Thu, 30 Aug 2001 21:48:47 -0400 (EDT)


Kurt Hornik <hornik at ci dot tuwien dot ac dot at> writes:
>
>> James R Van Zandt writes:
>
>> What is the relationship between the distribution calculated by
>> Octave's function and the tabulated distribution?  I'd like to add
>> that to the documentation.
>
>The Kolmogorov-Smirnov CDF corresponds to Dn*, i.e. the two-sided test.
>One could extend that to use 1 - exp(-2*x^2) for the one-sided tests,
>but I would use that for the tests but not the CDF.

But the usual Kolmogorov-Smirnov table still has two parameters, while
the Octave function has only one.  For example, for n = 7 samples and
a probability of .20 the table shows a critical value of Dn* = .381 .
Can I use the Octave function to calculate that?

	     - Jim Van Zandt