On Wed, Apr 23, 2014 at 12:57 AM, Gael Varoquaux
<gael.varoqu...@normalesup.org> wrote:
> On Tue, Apr 22, 2014 at 08:11:30PM -0700, Faraz Mirzaei wrote:
>> I'm putting together a regression test for a LogisticRegression classifier.
>> After fixing random_state, I get identical results on linux. But the same
>> random_state leads to a different result in windows.
>
> This is not entirely surprising. Controlling random_state (ie the seed of
> the random number generator) is a weak control of the set of random
> numbers drawn that will vary based upon architecture.

In case its relevant:
numpy random numbers are supposed to be identical across computers and versions.
Unit tests were added for this to numpy a few years ago.

Josef

>
> The question is rather: if you are getting vastly different coef_, you
> problem is probably very ill-posed. Are you getting np.dot(X, coef_) that
> varies widely or not? If it is not the case, it means that you are
> getting variations in the kernel of X, which isn't entirely surprising.
> If it is the case, I think you just need more regularization.
>
> Gaƫl
>
> ------------------------------------------------------------------------------
> Start Your Social Network Today - Download eXo Platform
> Build your Enterprise Intranet with eXo Platform Software
> Java Based Open Source Intranet - Social, Extensible, Cloud Ready
> Get Started Now And Turn Your Intranet Into A Collaboration Platform
> http://p.sf.net/sfu/ExoPlatform
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

------------------------------------------------------------------------------
Start Your Social Network Today - Download eXo Platform
Build your Enterprise Intranet with eXo Platform Software
Java Based Open Source Intranet - Social, Extensible, Cloud Ready
Get Started Now And Turn Your Intranet Into A Collaboration Platform
http://p.sf.net/sfu/ExoPlatform
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to