Hi Immanuel.
Sorry, I couldn't find my scripts. I'll look at home, though it is
possible I deleted them.
Have you tried dense data? Not sure how the sparse format works.
Have you uploaded something already?
I can't really help you today, but maybe tomorrow.
If you want, you can send me you hdf5 file.
Cheers,
Andy
Am 14.06.2012 13:23, schrieb iBayer:
> Hey Andreas,
>
> I don't get the hdf5 files though the mldata.org parser and the
> documentation is somewhat
> sparse... Do you have some old code laying around for setting up the hdf5
> files?
>
> Currently I'm using: ( and failing... )
> -----
> f = h5py.File('InternetAd.h5', 'w')
>
> # HDF5 attributes on the root level
> f.attrs['mldata'] = 0
> f.attrs['name'] = 'InternetAd'
> f.attrs['comment'] = 'Internet-Ad [Kushmerick, 1999]: document
> classification problem with \
> mostly binary features.'
>
>
> #HDF5 group 'data_descr'
> f.create_group('/data_descr')
> f['/data_descr/ordering'] = np.array(['label', 'data'])
>
> #HDF5 group 'data'
> f.create_group('/data')
> f['/data/data'] = X.data
> f['/data/data_indices'] = X.tocsc().indices
> f['/data/data_indptr'] = X.tocsc().indptr
> f['/data/label'] = y
>
> f.close()
>
> ------
> might have something to do with the ordering
>
>
>
> 2012/6/13 Andreas Mueller<[email protected]>:
>> Hi Immanuel.
>> When I worked on that, I used hdf5 files. I think
>> it is the best way to communicate with mldata.
>> Maybe try h5py.
>> Documentation for the mldata interface is here.
>>
>> You should not include the training/test splits in the datasets, but rather
>> create a "task" for that. Including the training/test splits will mess
>> with the automatic conversions.
>>
>> Hope that helps.
>>
>> Cheers,
>> Andy
>>
>> Am 13.06.2012 14:11, schrieb Immanuel B:
>>
>> Hello,
>>
>> I was going to upload some of the data sets listed here:
>> https://github.com/scikit-learn/scikit-learn/wiki/Setting-up-tests-to-benchmark-current-and-future-code
>> to mldata.org so make them easily available in scikit-learn.
>> The problem is that I can't find much information on how mldata.org
>> parses the uploaded data or how to make data available in different
>> formats.
>> So I just uploaded the data in the supported RData format, but that
>> didn't do much good.
>>
>> I also had a look at mldata-util (http://mloss.org/software/view/262/
>> )
>> but couldn't find any useful documentation either.
>>
>> Does someone know how to upload the data so that it can be retrieved
>> using fetch_mldata() ?
>>
>> Thanks,
>> Immanuel
>>
>> ------------------------------------------------------------------------------
>> Live Security Virtual Conference
>> Exclusive live event will cover all the ways today's security and
>> threat landscape has changed and how IT managers can respond. Discussions
>> will include endpoint security, mobile security and the latest in malware
>> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>> _______________________________________________
>> Scikit-learn-general mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Live Security Virtual Conference
>> Exclusive live event will cover all the ways today's security and
>> threat landscape has changed and how IT managers can respond. Discussions
>> will include endpoint security, mobile security and the latest in malware
>> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>> _______________________________________________
>> Scikit-learn-general mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
