Andreas Mueller <amueller@...> writes:

> 
> On 12/03/2012 09:39 PM, Afik Cohen wrote:
> > No, we aren't doing multi-label classification, just multiclass. He was 
saying
> > we could just use SGDClassifier directly, which is true, but AFAIK there is 
no
> > way to get good prediction probability outputs on a per-class basis unless 
you
> > train binary classifiers by wrapping it in a OneVsRestClassifier() call... 
or
> > is there?
> Hm it seems that the predict_proba is not implemented for multi-class.
> I thought someone had done that already, sorry.
> 
> It does train OVR and you can just take the sigmoid of the decision function
> and than normalize.
> Not sure why that is not implemented, I think we could basically
> copy and paste that from logistic regression.
> 


Will this let us run SGDClassifier and show us per-class probability outputs? 
Again, that's the only reason we've been using OneVsRestClassifier. Let me 
explain what I mean by per-class probability, just in case it isn't clear:

SGDClassifier's predict_proba() returns probability of belonging to each class,
so if for example there are five classes, it will return something like
[0.5, 0.3, 0.1, 0.05, 0.05]. 
For our use case, though, we need a negative/positive probability display, i.e.
[(0.4, 0.5), (0.7, 0.3), (0.8, 0.2), (0.9, 0.1), (0.6, 0.4)] for five classes
showing the probability that the input does not belong/does belong to that 
class, respectively.

Thanks, Afik



------------------------------------------------------------------------------
LogMeIn Rescue: Anywhere, Anytime Remote support for IT. Free Trial
Remotely access PCs and mobile devices and provide instant support
Improve your efficiency, and focus on delivering more value-add services
Discover what IT Professionals Know. Rescue delivers
http://p.sf.net/sfu/logmein_12329d2d
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to