[SciPy-user] new Kolmogorov-Smirnov 2-sample test not Numerical Recipes based

Robert Kern robert.kern@gmail....
Wed Dec 3 22:18:09 CST 2008


On Wed, Dec 3, 2008 at 20:58, Matthew Brett <matthew.brett@gmail.com> wrote:
> Hi,
>
>> I did a 2sample kstest based on the definition using search sorted,
>> see function below.
>
> [snip]
>
>> If there is any interest, I can replace the existing implementation
>> with the new function
>
> [snip]
>
>>  * in n1=2, n2=3 example new version is correct, old version is wrong
>>  * Monte Carlo essentially identical results (sample sizes between 100 and 1000
>>    rejection rates for alpha = 1%,5%,10% sometimes differ in 3rd decimal
>
> Given more secure license compatibility and that you think the current
> version is correct where the other is wrong, I think this should go in
> now, and into the 0.7 release.  Any disagreement?

None from me.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless
enigma that is made terrible by our own mad attempt to interpret it as
though it had an underlying truth."
  -- Umberto Eco


More information about the SciPy-user mailing list