From: Peter S. <pet...@gm...> - 2011-11-01 07:37:16
|
Hi, You can leave the response function on the default. I don't expect your result yo depend on this choice. There is no simple rule of which function is the best. It will influence the training behaviour but not very much the final result. The options which are more interesting for you to play with are ConvergenceTests ConvergenceImprove The sampling options and of course the architecture. Cheers, Peter d....@ii... schrieb: Dear TMVA experts, Is there any rule when to set the neuron response function to tanh or sigmoid or radial.Or I have to test which one gives the best response? Deepanwita Dutta Research Scholar Department of Physics IIT Guwahati Guwahati-781039 _____________________________________________ RSA® Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 _____________________________________________ TMVA-users mailing list TMV...@li... https://lists.sourceforge.net/lists/listinfo/tmva-users |