Thank you for your kind advice. I find a libSVM wrapper in scikit-learn and
wonder is that OK building a python wrapper for libFM using cython as a proper
contribution to scikit-learn?
some details:
libFM is a software implementation for factorization machines.you can find more
information here: http://www.libfm.org/
> From: [email protected]
> Subject: Scikit-learn-general Digest, Vol 39, Issue 13
> To: [email protected]
> Date: Mon, 8 Apr 2013 03:07:49 +0000
>
> Send Scikit-learn-general mailing list submissions to
> [email protected]
>
> To subscribe or unsubscribe via the World Wide Web, visit
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
> or, via email, send a message with subject or body 'help' to
> [email protected]
>
> You can reach the person managing the list at
> [email protected]
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Scikit-learn-general digest..."
>
>
> Today's Topics:
>
> 1. Re: Extend Python with Go? (Robert Layton)
> 2. Re: Extend Python with Go? (Mathieu Blondel)
> 3. Re: Use SciPy optimization instead of brute force GridSearch
> (Joel Nothman)
> 4. a questoin for scikit-learn's GSoC 2013 (YanChunwei)
> 5. Re: a questoin for scikit-learn's GSoC 2013 (Vlad Niculae)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 8 Apr 2013 10:21:23 +1000
> From: Robert Layton <[email protected]>
> Subject: Re: [Scikit-learn-general] Extend Python with Go?
> To: [email protected]
> Message-ID:
> <calcd8sgm4oredkfbttyybmq6c37tkq-zwyc+pg0920a2zwf...@mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Great :) Thanks.
>
>
> On 8 April 2013 09:23, Andreas Mueller <[email protected]> wrote:
>
> > On 04/08/2013 12:52 AM, Robert Layton wrote:
> > >
> > > Profiling can help with the optimisation, but it would be nice to have
> > > a pep8 style "warning line 32: missed type declaration" tool.
> > >
> > There is: cython -a
> >
> >
> >
> > ------------------------------------------------------------------------------
> > Minimize network downtime and maximize team effectiveness.
> > Reduce network management and security costs.Learn how to hire
> > the most talented Cisco Certified professionals. Visit the
> > Employer Resources Portal
> > http://www.cisco.com/web/learning/employer_resources/index.html
> > _______________________________________________
> > Scikit-learn-general mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
> >
>
>
>
> --
>
> Public key at: http://pgp.mit.edu/ Search for this email address and select
> the key from "2011-08-19" (key id: 54BA8735)
> -------------- next part --------------
> An HTML attachment was scrubbed...
>
> ------------------------------
>
> Message: 2
> Date: Mon, 8 Apr 2013 10:25:19 +0900
> From: Mathieu Blondel <[email protected]>
> Subject: Re: [Scikit-learn-general] Extend Python with Go?
> To: [email protected]
> Message-ID:
> <CAOKSrLzAdBJ0YQexKyMgmucmL6rF7hwH315ux=jgz9xmhre...@mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> On Mon, Apr 8, 2013 at 7:30 AM, Roman Sinayev <[email protected]> wrote:
>
> > I saw this http://gopy.qur.me/extensions/examples.html on hacker news
> > recently. It seems a lot cleaner than Cython. What do you guys
> > think?
> >
>
> It seems much more low-level to me than Cython: the user has to deal with
> reference counting...
>
> Mathieu
> -------------- next part --------------
> An HTML attachment was scrubbed...
>
> ------------------------------
>
> Message: 3
> Date: Mon, 8 Apr 2013 11:27:28 +1000
> From: Joel Nothman <[email protected]>
> Subject: Re: [Scikit-learn-general] Use SciPy optimization instead of
> brute force GridSearch
> To: [email protected]
> Message-ID:
> <CAAkaFLUXjKWx=WwYF1eRMW29gAvuwqw=BHWVMf1=p_ecrnk...@mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Currently BaseSearchCV expects a predetermined sequence of parameter
> settings, which is unideal for some cases. SciPy opts for a callback
> approach. I've not used that interface, but I gather something like this
> might work:
>
> class MinimizeCV(BaseEstimator):
> def __init__(self, minimiser, clf, param_init, scoring, cv,
> minimise_kwargs={}):
> self.clf = clf
> self.param_init = param_init
> self.scoring = scoring
> self.cv = cv
> self.minimiser = minimiser
> self.minimise_kwargs = minimise_kwargs
>
> def fit(self, X, y=None):
> def objective(param_values):
> """"""
> # TODO: parallelise fold fitting
> return aggregate(
> fit_grid_point(X, y, self.clf, dict(zip(param_list,
> param_values)), self.scoring, ...)
> for train, test in self.cv)
> res = self.minimiser(objective, X0=[v for k, v in
> sorted(self.param_init.iteritems())], **self.minimise_kwargs)
> # TODO: store results and perhaps search history
>
> I think a variant could be implemented that shares most of its code with
> the existing BaseSearchCV.
>
> I haven't looked at hyperopt's interface yet.
>
> - Joel
>
> > On Sun, Apr 7, 2013 at 6:35 PM, Roman Sinayev <[email protected]> wrote:
>
> > >
> > > It seems like brute force Grid Search takes forever when attempting to
> > > determine best parameters with many classifiers. Let's say the
> > > parameter space looks something like this
> > > http://i.imgur.com/AiBl8Wt.png . Why not use the SciPy simulated
> > > annealing or some simple genetic algorithm instead of searching
> > > through all the possible parameter space of every classifier?
> >
> -------------- next part --------------
> An HTML attachment was scrubbed...
>
> ------------------------------
>
> Message: 4
> Date: Mon, 8 Apr 2013 02:40:45 +0000
> From: YanChunwei <[email protected]>
> Subject: [Scikit-learn-general] a questoin for scikit-learn's GSoC
> 2013
> To: "[email protected]"
> <[email protected]>
> Message-ID: <[email protected]>
> Content-Type: text/plain; charset="gb2312"
>
> Hello, I am a computer science graduate student?and want to apply for
> scikit-learn's GSoC2013.I have a question, are all scipy-learn's tasks for
> GSOC's applicants the topics listed on the project's wiki on github?
> https://github.com/scikit-learn/scikit-learn/wiki/A-list-of-topics-for-a-Google-Summer-of-Code-%28GSOC%29-2013?
>
> I list some topics here:Add scipy.sparse matrix input support to the Decision
> Tree ImplementationOnline Low Rank Matrix CompletionOnline Non Negative
> Matrix Factorization... ...
> Is that possible for an applicant to implement other ideas?
> I have an idea, to implement a new algorithm for scikit-learn, the
> Feature-Based Matrix Factorization (FBMF). It is powerful and should have a
> long-term application in the field of Collaborative Filter.
> Similar to Factorization Machine , this model is an abstract of many
> variants of matrix factorization models, and new types of information can be
> utilized by simply defining new features, without modifying any lines of
> code.
> To build an open source implementation and integrate it into the
> scikit-learn, and use the rich algorithms provided by scipy-learn to
> automatically extract features, using python to control the overall training
> process (automatically grid search and so on), I think that is a wonderful
> choice.
> Is there anyone else think it is an good idea, and provide some suggestions
> for me to participate in GSOC of scipy-learn, I really appreciate it.
> Chunwei Yan
> -------------- next part --------------
> An HTML attachment was scrubbed...
>
> ------------------------------
>
> Message: 5
> Date: Mon, 8 Apr 2013 12:07:36 +0900
> From: Vlad Niculae <[email protected]>
> Subject: Re: [Scikit-learn-general] a questoin for scikit-learn's GSoC
> 2013
> To: "[email protected]"
> <[email protected]>
> Message-ID:
> <cafjw_ehaq7phq42j5u7ugs+932j09uh5w_cutyu1xc5a1nn...@mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
> Hello!
>
> In my opinion the ideas page needs a bit of restructuring, as the
> breadth and complexity of the suggested projects varies.
>
> We are definitely welcome to different ideas, as long as they prove
> useful and fit in the big picture for where the scikit is heading.
> This comes down to discussing it on the mailing list, so you did the
> right thing!
>
> I am personally not familiar with the algorithm you are talking about,
> could you please provide a reference paper? We have a semi-strict
> guideline on only considering algorithms that have many citations; I
> don't recall at the moment how many does "many" mean.
>
> Again, I don't know the algorithm, but I tend to think that a single
> algorithm is a contribution that is too small for a whole summer. I
> also think it would be better to have the standard SGD matrix
> factorization algorithm before implementing (what I suppose are)
> variants of it. But I don't want to discuss this further before
> looking
> at the paper, since I'm not standing on solid ground.
>
> I am excited for your interest in this. Remember, the first step, no
> matter what will be the project you would like to propose, is to start
> contributing one or several minor improvements or fixes, in order to
> get familiar with the codebase and with the way our development
> process works.
>
> Cheers,
> Vlad
>
>
> On Mon, Apr 8, 2013 at 11:40 AM, YanChunwei <[email protected]> wrote:
> > Hello, I am a computer science graduate student?and want to apply for
> > scikit-learn's GSoC2013.
> > I have a question, are all scipy-learn's tasks for GSOC's applicants the
> > topics listed on the project's wiki on github?
> > https://github.com/scikit-learn/scikit-learn/wiki/A-list-of-topics-for-a-Google-Summer-of-Code-%28GSOC%29-2013?
> >
> > I list some topics here:
> >
> > Add scipy.sparse matrix input support to the Decision Tree Implementation
> > Online Low Rank Matrix Completion
> > Online Non Negative Matrix Factorization
> > ... ...
> >
> >
> > Is that possible for an applicant to implement other ideas?
> >
> > I have an idea, to implement a new algorithm for scikit-learn, the
> > Feature-Based Matrix Factorization (FBMF). It is powerful and should have
> > a long-term application in the field of Collaborative Filter.
> >
> > Similar to Factorization Machine , this model is an abstract of many
> > variants of matrix factorization models, and new types of information can
> > be utilized by simply defining new features, without modifying any lines of
> > code.
> >
> > To build an open source implementation and integrate it into the
> > scikit-learn, and use the rich algorithms provided by scipy-learn to
> > automatically extract features, using python to control the overall
> > training process (automatically grid search and so on), I think that is a
> > wonderful choice.
> >
> > Is there anyone else think it is an good idea, and provide some suggestions
> > for me to participate in GSOC of scipy-learn, I really appreciate it.
> >
> > Chunwei Yan
> >
> > ------------------------------------------------------------------------------
> > Minimize network downtime and maximize team effectiveness.
> > Reduce network management and security costs.Learn how to hire
> > the most talented Cisco Certified professionals. Visit the
> > Employer Resources Portal
> > http://www.cisco.com/web/learning/employer_resources/index.html
> > _______________________________________________
> > Scikit-learn-general mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
> >
>
>
>
> ------------------------------
>
> ------------------------------------------------------------------------------
> Minimize network downtime and maximize team effectiveness.
> Reduce network management and security costs.Learn how to hire
> the most talented Cisco Certified professionals. Visit the
> Employer Resources Portal
> http://www.cisco.com/web/learning/employer_resources/index.html
>
> ------------------------------
>
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
> End of Scikit-learn-general Digest, Vol 39, Issue 13
> ****************************************************
------------------------------------------------------------------------------
Minimize network downtime and maximize team effectiveness.
Reduce network management and security costs.Learn how to hire
the most talented Cisco Certified professionals. Visit the
Employer Resources Portal
http://www.cisco.com/web/learning/employer_resources/index.html
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
