2011/12/5 Andreas Mueller <[email protected]>:
> On 12/05/2011 11:14 PM, Alexandre Gramfort wrote:
>>> I do not understand. I have the dictionary already, so what is being 
>>> estimated?
>> well I am not sure to follow now, but if you have the dictionary the
>> only missing part is the coefs of the decomposition.
>>
>> X = dico x coefs
> I think there is a little misunderstanding here.
> As I understand Ian, he has estimated a dictionary on some dataset
> and wants to use this dictionary to encode some "new" data.
>
> You do not need to estimate anything to get the "model".
>
> What you do want is to "transform" the new data so that it
> is coded using the specified dictionary.

Yes.

> I think this is exactly what the sparse encoding method that Olivier
> referenced is doing.

Yes and this method is just a wrapper for what Alexandre is explaining
(about LARS and OMP). Both replies are consistent.

Sparse coding is not simple linear projection (as in the transform
method of PCA for instance) as there is a penalty on the non-zero
loadings. Hence there is some "estimator fitting" that still occur at
the "transform" time. Unless of course who use fixed arbitrary
thresholding that might be enough for some tasks (e.g. sparse feature
extraction for image classification).

-- 
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel

------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure 
contains a definitive record of customers, application performance, 
security threats, fraudulent activity, and more. Splunk takes this 
data and makes sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-novd2d
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to