>
> Virgile? I thought that you had almost isolated a simple test case.
>
Not really, I could have the graph_lasso crash but I was using a
configuration in which my n_features-dimensional observations actually lied
in a (n_features - 1)-dimensional subspace.
Currently, I am using the graph_lasso to estimate covariance matrix on n x
p datasets with p / n > 1 (and matrix condition number up to 1000). I get a
lot a warning in the GraphLassoCV, but it can complete the computation
anyway. Whenever I get an exception, I catch it, ignore it, and rely on the
number of simulations that I perform to compensate the failure.
We may still find cases that yield an error, but it has not been a problem
anymore for me.
I guess I will go back to that issue when I work with real data.
On Wed, Dec 7, 2011 at 9:41 AM, Gael Varoquaux <
[email protected]> wrote:
> On Wed, Dec 07, 2011 at 09:38:32AM +0100, Olivier Grisel wrote:
> > > Can people confirm that some other solvers (e.g. GLMnet) do not have
> the
> > > same problem? In which case, we need to figure out how they do it.
>
> > Are we still talking about the Lasso with Coordinate Descent
>
> Yes.
>
> > Would be great to have a reproduction case in a gist
>
> Virgile? I thought that you had almost isolated a simple test case.
>
> G
>
------------------------------------------------------------------------------
Cloud Services Checklist: Pricing and Packaging Optimization
This white paper is intended to serve as a reference, checklist and point of
discussion for anyone considering optimizing the pricing and packaging model
of a cloud services business. Read Now!
http://www.accelacomm.com/jaw/sfnl/114/51491232/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
