> Y=None in pls._PLS, pls.PLSSVD

I might be wrong, but this is actually just a normal `y=None`. It's
capitalized because that's standard notation for PLS and it emphasises
the multi-target nature of the algorithm.  I tried changing it to
lowercase `y` once but I think we decided not to break the api for no
reason.  Maybe it's a good time to discuss it again.  Or maybe I
forget what the reason not to change it was.

Cheers,
Vlad

On Thu, May 16, 2013 at 11:03 PM, Lars Buitinck <[email protected]> wrote:
> 2013/5/16 Joel Nothman <[email protected]>:
>> I notice that there are some remnants of a former scikit-learn that allowed
>> parameters to fit(...) other than X and y, but these have been removed in
>> favour of object-level parameters.
>
> That's right. The old convention was that fit would call _set_params,
> then do actual training. The problem with this was that the method
> could fail halfway through, without signaling to the caller what
> exactly had failed, so we made set_params public and deprecated
> passing estimator parameters to fit.
>
>> Should the same be the case with transform? In particular, Pipeline and
>> FeatureUnion only support 1-ary transform, and any parameter that can only
>> be set as an additional argument to transform can't be varied in a [grid]
>> search. That is, >1-ary transform seems misfit to the API.
>
> Yes, preferably the same should be the case for transform.
>
>> So what non-X arguments to transform methods take?
>>
>> copy=True in various, such as feature_extraction.text.TfidfTransformer
>> y=None in various
>> ridge_alpha=None in decomposition.sparse_pca.SparsePCA
>> threshold=None in feature_selection.selector_mixin.SelectorMixin
>> pooling_func=np.mean in
>> cluster._feature_agglomeration.AgglomerationTransform
>> Y=None in pls._PLS, pls.PLSSVD
>>
>> copy is not really a problem. It's not something you want to vary in
>> parameter search, and although perhaps the pipeline could take advantage of
>> it, it's no problem.
>
> [Aside: a consistent way of telling transformers that they own
> whatever is passed to them could benefit pipeline performance, though
> I never took the time to design an API for that. Putting fake copy=
> arguments on all constructors or transform methods is ugly, this
> should be handled in TransformerMixin.]
>
> --
> Lars Buitinck
> Scientific programmer, ILPS
> University of Amsterdam
>
> ------------------------------------------------------------------------------
> AlienVault Unified Security Management (USM) platform delivers complete
> security visibility with the essential security capabilities. Easily and
> efficiently configure, manage, and operate all of your security controls
> from a single console and one unified framework. Download a free trial.
> http://p.sf.net/sfu/alienvault_d2d
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

------------------------------------------------------------------------------
AlienVault Unified Security Management (USM) platform delivers complete
security visibility with the essential security capabilities. Easily and
efficiently configure, manage, and operate all of your security controls
from a single console and one unified framework. Download a free trial.
http://p.sf.net/sfu/alienvault_d2d
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to