Greetings all!
Sorry for the delay. Thanks for all your answers !
Vlad, regarding the multi-class (orr multi-label) options, as fas as I
know, there is not a lot of documentation on the topic. I did once
implemented the multi-class option for the project I was working on, in an
OvR fashion. In the code I put on the Gist there is just the binary option
cause I wanted to keep it well (or better) documented. The multi-label
option would be even less documented I guess. Anyway, I am currently
looking into that.
About having the transformers separated, yes I agree. In fact I started
coding it that way but I read the docstring of the TfidfTransformer and it
mentions the SMART notation, so I thought it would make sense to have all
the SMART options in the same transformer. Unfortunately, it's true! It is
a Christmas-tree-type class already, and the arguments may contradict each
other so it is kinda messy.
I will work then on the BM25 separate transformer, with tests (which I am
quite sure I will be asking about) and documentation. I will follow the
contributing guidelines to make it easier to see what am I working on. For
DeltaIdf, I guess it is better to wait for Lars tf-chi2 PR, and then see
whether it's worthwhile or not.
I will read the papers, specially about supervised weighting to make a
better case for them and their usefulness.
Cheers guys,
Pavel
On Mon, Aug 25, 2014 at 2:08 PM, Lars Buitinck <[email protected]> wrote:
> 2014-08-23 17:06 GMT+02:00 Lars Buitinck <[email protected]>:
> > [3] This paper from a guy at HP Research that I cannot find right now.
>
> Found it: Forman et al., Feature Shaping for Linear SVM Classifiers,
> http://www.hpl.hp.com/techreports/2009/HPL-2009-31R1.pdf (SIGKDD
> 2009).
>
>
> ------------------------------------------------------------------------------
> Slashdot TV.
> Video for Nerds. Stuff that matters.
> http://tv.slashdot.org/
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
