7

I have a very general question: how do I choose the right kernel function for SVM? I know the ultimate answer is try all the kernels, do out-of-sample validation, and pick the one with best classification result. But other than that, is there any guideline of trying the different kernel functions?

BenMorel
  • 34,448
  • 50
  • 182
  • 322
CuriousMind
  • 15,168
  • 20
  • 82
  • 120
  • 1
    Most of the times the best kernel to use is RBF kernel. The reason is most of the things follow Gaussian distribution. The answer of "why" is related to the [central limit theorem](http://en.wikipedia.org/wiki/Central_limit_theorem) according to me. – Autonomous Jun 06 '13 at 14:44

2 Answers2

5

Always try the linear kernel first, simply because it's so much faster and can yield great results in many cases (specifically high dimensional problems).

If the linear kernel fails, in general your best bet is an RBF kernel. They are known to perform very well on a large variety of problems.

Marc Claesen
  • 16,778
  • 6
  • 27
  • 62
  • How do you know if the linear kernel "fails"? Is it if the linear kernel model performs worse than other models besides SVM? – VMS Aug 16 '17 at 04:46
2

Look here to find the answer.

https://stats.stackexchange.com/questions/18030/how-to-select-kernel-for-svm

Basically, there is rather no one good path to choose, unless you know something important about your data that might determine proper kernel to use. However, follow the link above to get more specific information.

Community
  • 1
  • 1
storaged
  • 1,837
  • 20
  • 34