From help-request at octave dot org Thu Nov 17 14:28:48 2005 Subject: Re: Kolmogorov-Smirnov test From: Mike Miller To: Hamish Allan cc: Brendan Drew , Help-Octave List Date: Thu, 17 Nov 2005 14:20:31 -0600 (CST) On Thu, 17 Nov 2005, Hamish Allan wrote: > On 17 Nov 2005, at 19:25, Mike Miller wrote: > >> In statistical testing, a valid p-value has a uniform distribution when >> the null hypothesis is true. In the K-S two-sample test, the null >> hypothesis is that the two distributions are the same. We reject the >> null when p is small. The probability of p < .05, for example, equals >> .05 when the null is true, and this is because p is uniform when the >> null is true. > > Right, I certainly didn't understand this, and now the results seem less > strange :) > > So is there any way of doing what I originally wanted to do: determine > how likely two given datasets are to have been drawn from the same > distribution? > > In fact, what I want to do is to try to determine how "representative" a > subsample is of the distribution of the original sample. This involves > plenty of ties... so should I be going about this a whole different way? > Any pointers gratefully accepted. The k-s test (not k-s2) is used for testing if a sample came from a given population. In your case the "given population" is given by the original sample. It tests a particular kind of deviation though, maybe not what you want to know. I don't know what it does with ties in the Octave version. Mike ------------------------------------------------------------- Octave is freely available under the terms of the GNU GPL. Octave's home on the web: http://www.octave.org How to fund new projects: http://www.octave.org/funding.html Subscription information: http://www.octave.org/archive.html -------------------------------------------------------------