CODEWITHSUNDEEP
machine-learningartificial-intelligencedata-miningsvm
London guy
London guy

Reputation: 28012

Good resource to learn about SVM paramaters in LibSVM

What is a good resource to learn about what are the various parameters in the LibSVM and how to use them? I have seen several discussions about the cost parameter and the gamma parameter, the cost weights and so on. I am unable to get the hang of these parameters because I do not know what they mean and how should they be used to improve my classifier under various situations.

Is there a comprehensive resource that explains these for SVM classification?

Upvotes: 1

Views: 376

Answers (1)

AGS
AGS

Reputation: 14498

The libsvm FAQ is helpful. As far as the C and gamma paramaters go, the C parameter is the cost parameter of the error term, used in the soft margin SVM. The gamma parameter is the single parameter in the gaussian RBF. Optimal values of both can be found using a simple grid search.

Upvotes: 2

Related Questions