Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

Loading…

Add multiclass support to hinge_loss #3451

Closed
arjoly opened this Issue · 6 comments

4 participants

@arjoly
Owner

The hinge_loss metrics could be improved by adding support multiclass hinge loss. For more information, see wikipedia and the narrative documentation.

This would require to enhance the current code, write tests and modify the documentation accordingly.

@mblondel
Owner

There are several multiclass variants of the hinge loss. The one in the article you link is the variant by Crammer & Singer. We should either accept a parameter to let the user choose or use one function per variant.

@mblondel
Owner

We can also implement the "one-vs-rest" loss, which is simply to sum the hinge losses of each binary classification problem.

@arjoly
Owner

Thanks @mblondel for the precision.

@SaurabhJha

I am new to Scikit-learn. I am trying to fix this issue.

@SaurabhJha

Hi, I am reading this paper
http://www.ttic.edu/sigml/symposium2011/papers/Moore+DeNero_Regularization.pdf
in order to understand the notation of multi class hinge loss equation at wikipedia(w subscript y and w subscript t). Can anyone please confirm if w subscript f and l are functional margins? Are they product of w evaulated at some feature and the corresponding product?

Also, is there any way to find w from est?

@MechCoder
Owner

Fixed by #3607

@MechCoder MechCoder closed this
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.