The vanilla rakel and vanilla classifier chain would be a great addition
in scikit-learn.
FYI
For the classifier chain, there is a stalled pull request
https://github.com/scikit-learn/scikit-learn/pull/3727
<https://github.com/scikit-learn/scikit-learn/pull/3727> .
For the rakel classifier, three features are needed:
1. multi-output support in the bagging estimator
2. A working label power set transformer and classifier
3. sub-sampling of the output (here label) space in the bagging
estimators
or in the label power set transformer
For multi-output bagging, there is already this pull request waiting for
review
https://github.com/scikit-learn/scikit-learn/pull/4848
<https://github.com/scikit-learn/scikit-learn/pull/4848> . This would also
enable
ensemble of classifier chain.
For the label power set, there is this stalled pull request
https://github.com/scikit-learn/scikit-learn/pull/2461
<https://github.com/scikit-learn/scikit-learn/pull/2461> .
Best regards,
Arnaud
> On 12 Jul 2015, at 20:37, Al <alain.pen...@gmail.com> wrote:
>
> Only 1 of the variants can be found in the litterature (with corrects
> parameters) as the random (ensemble) classifier chain.
> Other variants can not be found in the litterature, but I used them to
> compare several order strategies for my thesis, and thus will probably
> be removed if it is integrated as they have no citation and are too recent.
>
> Should I then begin to further improve the classifier chain to make it
> closer to the litterature (removing the variants, eventually removing
> the flexibility concerning the steps, ...)? And rakel?
>
>> I think classifier chains would be a nice addition.
>> I am not very familiar with the different variants you implemented, though.
>>
>>
>> On 07/10/2015 09:32 AM, Al wrote:
>>> I should probably have put a link to the current implementation as it is
>>> now. The link to this project (purged, i removed the various things from
>>> my thesis) is:
>>>
>>> https://github.com/Al-Pena/ClassifierChain
>>>
>>>> 2015-07-10 15:20 GMT+02:00 Al <alain.pen...@gmail.com>:
>>>>> My name is Alain Pena, (now previously) student in computer engineering
>>>>> at University of Liège.
>>>>> For my master thesis, I had to implement some methods for multilabel
>>>>> classification, those methods being RAKEL [1] and (Ensemble) Classifier
>>>>> Chain [2], as well as some variants of this latter (order of the chain
>>>>> or length of its links for example).
>>>>>
>>>>> They are currently lazy (I had a problem with memory while doing my
>>>>> thesis, so I had to implement them lazily, throwing each estimator away)
>>>>> as well as single threaded. They are tested for multilabel only with a
>>>>> test coverage of about 80%.
>>>>>
>>>>> Before eventually upgrading them to make them more robust and versatile,
>>>>> I wondered if scikit-learn would have any interest in those methods.
>>>> I've used a (hacky) homegrown RAKEL with some success [1], and I'm
>>>> interested to see the code and maybe get it in sklearn eventually.
>>>> Publishing as a separate project or a bunch of gists [2] would be a
>>>> good idea. I never used classifier chains, but I know the idea.
>>>>
>>>> [1]
>>>> https://staff.fnwi.uva.nl/m.derijke/wp-content/papercite-data/pdf/buitinck-multi-emotion-2015.pdf
>>>> [2] https://gist.github.com
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Don't Limit Your Business. Reach for the Cloud.
>>>> GigeNET's Cloud Solutions provide you with the tools and support that
>>>> you need to offload your IT needs and focus on growing your business.
>>>> Configured For All Businesses. Start Your Cloud Today.
>>>> https://www.gigenetcloud.com/
>>>> _______________________________________________
>>>> Scikit-learn-general mailing list
>>>> Scikit-learn-general@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>> ------------------------------------------------------------------------------
>>> Don't Limit Your Business. Reach for the Cloud.
>>> GigeNET's Cloud Solutions provide you with the tools and support that
>>> you need to offload your IT needs and focus on growing your business.
>>> Configured For All Businesses. Start Your Cloud Today.
>>> https://www.gigenetcloud.com/
>>> _______________________________________________
>>> Scikit-learn-general mailing list
>>> Scikit-learn-general@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>> ------------------------------------------------------------------------------
>> Don't Limit Your Business. Reach for the Cloud.
>> GigeNET's Cloud Solutions provide you with the tools and support that
>> you need to offload your IT needs and focus on growing your business.
>> Configured For All Businesses. Start Your Cloud Today.
>> https://www.gigenetcloud.com/
>> _______________________________________________
>> Scikit-learn-general mailing list
>> Scikit-learn-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
> ------------------------------------------------------------------------------
> Don't Limit Your Business. Reach for the Cloud.
> GigeNET's Cloud Solutions provide you with the tools and support that
> you need to offload your IT needs and focus on growing your business.
> Configured For All Businesses. Start Your Cloud Today.
> https://www.gigenetcloud.com/
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
