Xapian devel - Mar 2012 - Wish To Join Xapian:-)

Dear Friends,

This is Shao from National University of Singapore(NUS). I'm currently
doing my exchange study in Royal Institute of Technology(KTH), Sweden. IR
is really interesting to me. I've taken a Information Retrieval course
during the exchange study here in KTH
http://www.csc.kth.se/utbildning/kth/kurser/DD2476/ir12/labblydelser/assignment2.
 The
Weighting Schemes and Learn to Rank projects in Xapian sound very relevant
and interesting.I'd like to try out these weighting schemes to compare the
speed and precision. May I know besides the traditional weight schemes, are
we going to take the relevance feedback into account at this stage of
Xapian development?

Now I am doing a Music Information Retrieval project
http://www.csc.kth.se/utbildning/kth/kurser/DD2476/ir12/pdf/project1.pdf with
four of my teammates. Amazingly, three of them are going to Google for a
summer intern. I'm not that lucky to become a Googler but I really look
forward to get a GSOC experience which is really cool, too. Wish to join
you soon:-)

-- 
Best Regards,
Shaohuan Li
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xapian.org/pipermail/xapian-devel/attachments/20120321/7ef65cf2/attachment.html>

Hi Shao,

On Wed, Mar 21, 2012 at 07:11:52PM +0100, Shaohuan Li
wrote:
> May I know besides the traditional weight schemes, are we going to
> take the relevance feedback into account at this stage of Xapian
> development?
That's a good question.

Currently the relevance feedback in Xapian always uses the formula which
is equivalent to "TradWeight" document weighting.  But it's
abstracted to
an ExpandWeight class internally, so plugging in other formulae wouldn't
be too hard if they are expressible in a suitable form.

I think it's up to you whether you want to tackle that as part of this
project or not.
> Now I am doing a Music Information Retrieval project
> http://www.csc.kth.se/utbildning/kth/kurser/DD2476/ir12/pdf/project1.pdf
with
> four of my teammates. Amazingly, three of them are going to Google for a
> summer intern. I'm not that lucky to become a Googler but I really look
> forward to get a GSOC experience which is really cool, too. Wish to join
> you soon:-)
Heh.

I've not come across the "Million Song Dataset" before - that
looks an
interesting resource:

http://labrosa.ee.columbia.edu/millionsong/

Cheers,
    Olly