status of SvmLightRank

[Miller, Timothy]

I'm interested in using svm ranking for a problem. It looks like there
are some classes that use svm rank but I'm curious whether they are in a
usable state and whether there might be any examples of them in use?

I thought they were ready but while I can see how one might use the
DataWriter (creating QidInstances and passing them to write()), I have
trouble seeing how it would work at test time (the SvmLightRank class
overrides Classifier, which provides a classfy() method but it doesn't
accept Instances, just List<Feature>'s. So I'm not sure how one is meant
to inform the classifier what the set of instances to compare is.
Please let me know if there is something I'm missing. If I can get it
working then I'd willing to contribute some examples for future users.

Thanks
Tim Miller

[Steven Bethard]

Tim and I talked about this in person, and I think the conclusion is that while SVM-rank is trained to optimize a pairwise metric, it still produces a linear model in the end ("SVMrank learns an unbiased linear classification rule (i.e. a rule w*x without explicit threshold" https://www.cs.cornell.edu/people/tj/svm_light/svm_rank.html) so you should be able to just run the classifier on each example in your group, and then sort them by the resulting classifier score. If anyone wants to dig into it further and confirm this, the paper to look at is http://www.joachims.org/publications/joachims_02c.pdf.

--
You received this message because you are subscribed to the Google Groups "cleartk-developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cleartk-develop...@googlegroups.com.
To post to this group, send email to cleartk-d...@googlegroups.com.
Visit this group at http://groups.google.com/group/cleartk-developers.
For more options, visit https://groups.google.com/d/optout.