There was a thread back in April about including some metric learning
algorithms in scikit-learn. (
http://www.mail-archive.com/scikit-learn-general@lists.sourceforge.net/msg06952.html
)
The consensus at that time was to come up with some metric learning code in
a separate project first, to get a
all or anyone else is still interested I'm going to get going on this a bit
> soon.
>
> - John
> -- next part --
> An HTML attachment was scrubbed...
>
> --
>
> Message: 3
> Date: Tue, 14 May 2013 22:58:29 -0700
&
Perhaps a separate github project might be easiest.. Less headachey than
doing it with one branch and others PR-ing to it. At least that was the
case when I tried.
2013/5/15 Robert McGibbon
> John,
>
> What's the best way to work on a collaborative project like this? I've
> never tried working
John,
What's the best way to work on a collaborative project like this? I've never
tried working on a PR w/ multiple authors in git. Do you want to set up a new
github project for this during development?
-Robert
On May 14, 2013, at 10:27 PM, John Collins wrote:
> Hi there,
>
> A few weeks
Hi there,
A few weeks ago I posted about this. I have been finishing my thesis and
working concurrently and have had no free time but now I have some to
commit to this. At that point Kenneth C. Arnold and Robert McGibbon
mentioned they were also interested. I've done a bit of translation of a
piec
That's also the way I've used these techniques in the past: Build the
matrix A. Transform the space X to Y = A^(1/2) X. Then apply via knn or whatever takes your fancy.
- John
--
Try New Relic Now & We'll Send You this Co
er of different projects in the SciPy
> ecosystem.
>
> http://scikits.appspot.com/
>
> The term is similar in scope to the MATLAB(TM) "Toolbox" term. You
> would not title a book about the MATLAB(TM) Signal Processing
> Toolbox(TM) _Learning Toolbox_.
>
> Good luck in your search.
>
> --
&
For Mahalanobis metric, maybe we can do a cholesky decomposition to the
learned metric and make it a transformer? Then after it we can chain a knn
classifier after the transform.
Best,
Wei
On Tue, Apr 23, 2013 at 3:59 PM, Robert McGibbon wrote:
> Input to such algorithms is usually given as:
>
> Input to such algorithms is usually given as:
> - a set of similarity and dissimilarity links,
> - relative comparisons (x is closer to y than w is to z), or
> - target distances (x should be no farther than q from y).
>
> The outputs of all methods I've worked with are Mahalanobis distanc
T
Some gists:
https://gist.github.com/kcarnold/5439917
https://gist.github.com/kcarnold/5439945
They are rather terribly documented, sorry.
Input to such algorithms is usually given as:
- a set of similarity and dissimilarity links,
- relative comparisons (x is closer to y than w is to z), or
-
On Mon, Apr 22, 2013 at 3:56 PM, Alexandre Gramfort <
alexandre.gramf...@inria.fr> wrote:
> > If people were interested in putting together a separate package in the
> > style of the scikit collecting metric learning algorithms with a common
> API,
> > I would love to contribute to that too.
>
I'
It seems like there is already a manifold learning project in progress.
These two topics are closely related.
http://www.cs.cmu.edu/~liuy/lle_isomap_metric.pdf
--
Precog is a next-generation analytics platform capable of a
Does anyone have any code for computing rotations of components after
> PCA or FactorAnalysis, etc. E.g., varimax?
>
> Thanks,
>
> Skipper
>
>
>
> --
>
> Message: 2
> Date: Sun, 21 Apr 2013 18:43:44 -0700
> From: Robert McGibbon
> Subject: Re: [Scikit-learn-general] Metric Le
> If people were interested in putting together a separate package in the
> style of the scikit collecting metric learning algorithms with a common API,
> I would love to contribute to that too.
such ideas should be added to the next sprint wiki page so it can be
discussed among the devs
A
-
it-learn-general
>
> ------ next part ----------
> An HTML attachment was scrubbed...
>
> --
>
> Message: 3
> Date: Sun, 21 Apr 2013 16:38:13 -0600
> From: Gael Varoquaux
> Subject: Re: [Scikit-learn-general] Random patches a
An HTML attachment was scrubbed...
>
> ------
>
> Message: 3
> Date: Sun, 21 Apr 2013 16:38:13 -0600
> From: Gael Varoquaux
> Subject: Re: [Scikit-learn-general] Random patches and coordinates
> To: scikit-learn-general@lists.sourceforge.net
> Mess
If people were interested in putting together a separate package in the style
of the scikit collecting metric learning algorithms with a common API, I would
love to contribute to that too.
-Robert
On Apr 21, 2013, at 3:35 PM, Robert McGibbon wrote:
> This would be AWESOME.
>
> I have code imp
This would be AWESOME.
I have code implementing Shen, C.; Kim, J.; Wang, L. Scalable large-margin
Mahalanobis distance metric learning. IEEE Trans. Neural Networks 2010, 21,
1524–1530, but it yeah, it's not up to sklearn standards either.
-Robert
On Apr 21, 2013, at 12:49 PM, Kenneth C. Arnol
I have implemented a few metric learning algorithms myself. The quality of
that code is nowhere near sklearn standards, but I may have some incentive
to improve it soon.
-Ken
On Sun, Apr 21, 2013 at 3:42 PM, John Collins wrote:
> Has anybody or does anybody have plans to implement metric learn
Has anybody or does anybody have plans to implement metric learning
algorithms like ITML in sklearn?
If not, I would like to consider working on this.
Thanks,
John
--
Precog is a next-generation analytics platform capabl
20 matches
Mail list logo
