On 01/25/2013 11:48 AM, Olivier Grisel wrote: > 2013/1/24 Osman Başkaya <osman.bask...@computer.org>: >> Thanks Olivier. But I want to ask a question about this. Isn't it a problem >> that we give such a big steps? >> >> I worked on Weka a bit more and it seems that Logistic Regression in scikit >> is similar to Weka's LibLINEAR. I tried them on the same dataset. Results >> are similar except especially accommodate. I gave the same parameter values >> (C = 191, tol=0.0001, L2-regularized logistic regression, no randomization, >> no normalization etc..) to both classifiers and Weka gives me 0.58, but LR >> in scikit gives me 0.46. >> >> I am tired a bit now :) > We recently fixed a bug in the liblinear wrappers. Please make sure you are running the current release, 0.13. LogisticRegression is also based on LibLinear, with the loss set to logloss. To get the hinge loss, you have to use LinearSVC, which is also based on LibLinear. Best, Andy
------------------------------------------------------------------------------ Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft MVPs and experts. ON SALE this month only -- learn more at: http://p.sf.net/sfu/learnnow-d2d _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
