Hi Antoine!
The Kolmogorov test is not more robust when parameters are estimated: Kolmogorov test is wrong when parameters are estimated from data.
The question has already been discussed at:
hello,
This might be a naive question but i was surprised to find that KS test in OT does not show a “deterministic” behavior. The resulting p-value is sometimes big enough, sometimes very close to 0, for a given sample and a given theoretical distributionFactory object.
am i missing something here ? or is it just some numerical approximation thing ?
thanks,
sanaa
and
opened 07:51PM - 03 Sep 20 UTC
doc
WHen using the Kolmogorov-Smirnov Fitting test in the case where the distributio… n belongs to a given factory butits parameters are unknown, the returned distribution is correct (good estimated parameters), as well as the returned KS statistic.
However, the returned P-value is bad: when running the script, the P-value seems to take random values but always with 1 digit after the comma (example sequence: 0.1 ; 0.4 ;0.2 ;0.2 ;0.3; 0...)
The example below is relative to the NormalFactory
[ajust_cin_bug_KS.txt](https://github.com/openturns/openturns/files/5170831/ajust_cin_bug_KS.txt)
This might be interesting to know for Persalys
Best regards,
Michaël