2011/12/29 Gael Varoquaux <[email protected]>:
> On Thu, Dec 29, 2011 at 12:46:36PM -0800, adnan rajper wrote:
>>    I use LinearSVC for text classification. My problem is that I want to
>>    generate ROC curve for LinearSVC. Since LinearSVC does not output
>>    probabilties. Is there any other way to  generate ROC curve for LinearSVC?
>>    I have tried svm.SVC(kernel='linear', probabilities=True) but it gets too
>>    slow.
>
> If you really want to use LinearSVC (you could use a LogisticRegression
> instead), I see a couple of ways:
>
>  * You could bootstrap the prediction, and thus get some decision metric.
>   This is quickly going to get costly.
>
>  * You could use the decision function, (decision_function method of the
>   LinearSVC) although this is clearly a hack.

Why is this a hack? ROC is only concerned with the relative positions
of the decision threshold, not the probability normalization AFAIK, am
I wrong?

-- 
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel

------------------------------------------------------------------------------
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
infrastructure or vast IT resources to deliver seamless, secure access to
virtual desktops. With this all-in-one solution, easily deploy virtual 
desktops for less than the cost of PCs and save 60% on VDI infrastructure 
costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to