I am newbie to using svm for classification. I want to tune svm parameters by .TrainAutofunction in EmguCV. But I don't know what are the range(min-max value) of below parameters that I should give to this function to search:
1- C (for poly and RBF kernels)
2- Gamma (for poly and RBF kernels)
3- Coeficient (for poly kernel)
4- Degree (for poly kernel)
What are the range of these parameters?
Do these ranges depends on the number of samples?
As I receive the out of memory allocation error, can I set step parameter to a large value and when I found optimal values approximately, set these range to a smaller one and search on the second range with smaller step?
In general, you should consider the RBF kernel first, unless you have a good reason not to. Here are some guides on how to choose the parameters C and gamma:
http://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf
http://scikit-learn.org/stable/auto_examples/svm/plot_rbf_parameters.html