On May 5, 9:57 am, William Cauchois <[email protected]> wrote:
> In my opinion SVM functionality could be a useful addition to Sage.

Me too. In my opinion the optional SPKG for your binding should only
be the first step. The ideal situation for the future is a SVM class
in the statistics section of sage, where you can pose an SVM problem
in general and then select which backend to use. Notice, that there
are already two emerging possibilities for SVMs in Sage: The first one
via R (if you install the "e1071" R package) and the second one is
part of scipy.learn: 
http://fseoane.net/blog/2010/plot-the-maximum-margin-hyperplane-with-scikitslearn/

In general the goal of sage is to combine such tools under a common
umbrella, just like it is currently developed for (mixed integer)
linear programming in optimization by introducing a common class to
pose such problems and a solve method, where you select the solver
engine.

H

-- 
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to