Hi Andy,
I'm actually thinking about 1d convolution. I'd like to do sparse coding
on 1D multivariate signals. So far I used my own code, but I was hoping
to find better optimized and well-maintained implementations in
scikit-learn.
I think it would be nice to have some 1d convolutional algorithms in
scikit learn, since they are heavily used in signal processing for
audio, bio, and movement signals. I would even give it a try on my own.
But I guess it would not be as simple as subclassing some extisting
implementations and do some minor modifications, since the code in
scikit-learn seems pretty much optimized. Or is it?
Am 28.09.2012 14:29, schrieb Andreas Mueller:
> Hi Christian.
> Are you thinking about 1d or 2d convolutions?
> I am not so familiar with 1d signal processing but there has
> been some work on convolutional sparse coding for image patches.
> This is not really planned for sklearn, afaik, though.
> In computer vision, I think there was no big difference in recognition
> performance between convolutionally learned codes and codes
> that where trained on patches and applied independently.
>
> Hth.
> Andy
>
> Am 28.09.2012 10:59, schrieb Christian Vollmer:
>> Hello,
>>
>> there is a nice collection of sparse coding and dictionary algorithms
>> implemented in scikit-learn. However, it seems there are no
>> shift-invariant implementations. Are there plans to include any
>> shift-invariant implementations or is there a way to apply the
>> implemented algorithms in a shift-invariant manner?
>>
>> With "shift-invariant" I mean a formulation, where the reconstruction is
>> done by convolution of the atoms with the coefficients, like
>> min_{U,V}||X-conv(U,W)||_2^2|| + alpha ||U||_1
>>
>> I think, I could get shift-invariant behaviour in the coding part by
>> first building a dictionary of all shifted versions of all atoms and
>> then apply the implemented sparse coding algorithms. However, I don't
>> see a shift-invariant way for the dictionary learning part.
>>
>> Thanks,
>> Christian
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Got visibility?
>> Most devs has no idea what their production app looks like.
>> Find out how fast your code is with AppDynamics Lite.
>> http://ad.doubleclick.net/clk;262219671;13503038;y?
>> http://info.appdynamics.com/FreeJavaPerformanceDownload.html
>> _______________________________________________
>> Scikit-learn-general mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
> ------------------------------------------------------------------------------
> Got visibility?
> Most devs has no idea what their production app looks like.
> Find out how fast your code is with AppDynamics Lite.
> http://ad.doubleclick.net/clk;262219671;13503038;y?
> http://info.appdynamics.com/FreeJavaPerformanceDownload.html
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
------------------------------------------------------------------------------
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
