Hello,
I was working on this issue
https://github.com/scikit-learn/scikit-learn/issues/2402 and had submitted
this PR https://github.com/scikit-learn/scikit-learn/pull/2590 .
Since in this PR, I tried doing it under ElasticNetCV itself, according to
the suggestions of agramfort, one has to cross-validate alpha per
task/output. This leaves self.alpha_ to be an array and in order to use
ElasticNet, I can use only one alpha. For now I have taken the alpha which
gives the lowest MSE across all tasks which IMO is completely wrong. (see
the XXX note in the PR)
I think one way to solve this problem is
1] To have a separate MultiTaskElasticNetCV / MultiTaskLassoCV (and)
2] Make both MultiTaskElasticNet and MultiTaskLasso accept alpha as a
array-like object of floats, so that ElasticNet need not be fit multiple
times for each task in the new MultiTaskElasticNetCV.
I would like to know your opinion.
--
Regards,
Manoj Kumar,
Mech Undergrad
http://manojbits.wordpress.com
------------------------------------------------------------------------------
DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps
OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access
Free app hosting. Or install the open source package on any LAMP server.
Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native!
http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
