Hi Philipp. Great to hear that someone is using that :) The problem is that the approximation uses a "log". Afaik even the exact kernel is not defined if two features are compared that are both exactly zeros. Usually I just work around that by adding an epsilon. I was considering adding that to the code. What do you think?
Cheers, Andy Am 30.05.2012 13:29, schrieb Philipp Singer: > Hey! > > I am currently using this kernel approximation method followed by a > linear SVC. It works pretty well, but today I hopped into a problem: > > It seems like all the feature values need to be strict positive (when I > look into the code exception>= 0). > > Why is this the case? Is it somehow possible also to work with zero values? > > Regards, > Philipp > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
