You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am very curious why LIBSVM is superior (isn't it? ). Since a general SVM problem is usually a quadratic program or semi-definite program, if I solve it using some other efficient general-purpose quadratic or semi-definite program solvers, will higher performance than LIBSVM be achieved? In other words, I wonder what more work LIBSVM does at the technical level than other QP/SDP solvers? (No offense, I've always been a big fan of LIBSVM.)
Thank you in advance!
The text was updated successfully, but these errors were encountered:
As I understand it, LIBSVM (the C++ library that this package calls) implements a few solvers that specifically target the SVM problems. I'd expect that a specific solver has some benefits over a general one, but I know almost nothing about the solvers.
Nevertheless, I am closing this issue because it is not really an issue. Please, create a topic at Discourse if you want to discuss your question with the julia community, or contact the original author at https://github.com/cjlin1/libsvm.
I am very curious why
LIBSVM
is superior (isn't it? ). Since a general SVM problem is usually a quadratic program or semi-definite program, if I solve it using some other efficient general-purpose quadratic or semi-definite program solvers, will higher performance thanLIBSVM
be achieved? In other words, I wonder what more workLIBSVM
does at the technical level than other QP/SDP solvers? (No offense, I've always been a big fan ofLIBSVM
.)Thank you in advance!
The text was updated successfully, but these errors were encountered: