[pymvpa] [Fwd: Re: top-n match lists]

[Tara Gilliam]

Sorry, forgot to include reply-all!

-------- Original Message --------
Subject: Re: [pymvpa] top-n match lists
Date: Thu, 07 Jan 2010 18:10:37 +0000
From: Tara Gilliam <tg at cs.york.ac.uk>
To: Yaroslav Halchenko <debian at onerussian.com>

Hi Yaroslav,

Thanks for the (always quick!) reply. I'm kicking myself for not having thought
of this when Michael sent out the email for v0.5 feature requests!

At the moment I'm using SMLR and kNN, but am hoping to include ENET and possibly
SVM in the future as well. For kNN, are the values the number of votes per
sample in the training set, or per class?

Thanks,
Tara

Yaroslav Halchenko wrote:
> Hi Tara,
> 
> Unfortunately ATM we simply do not have such feature unified and
> exposed at all, but it sounds like a nice feature to have!  Probably
> something like 'predictions_ranked'  where for each prediction it would
> not only store the winner but all labels ordered accordingly.
> 
> I have filed the request under
> http://github.com/hanke/PyMVPA/issues#issue/8
> 
> Meanwhile you can make use of 'values' state variable.  The difficulty
> might be that the content of values is not unified across the
> classifiers, but it is somewhat obvious for most (e.g. for SMLR it would
> be probabilities from logistic regressions per each label in the order
> of your ds.uniquelabels, so you would need just to argsort each one of
> them with [::-1] to get in reverse order ;) ). What classifier do you
> have in mind though?
>