best accuracy from 10-fold Cross validation on SVM with grid search optimal gamma value and C

Kian

How i find the best accuracy from 10-fold Cross validation on SVM with grid search for finding optimal  gamma value and C?

how i can set the operators in Rapidminer for finding the optimal value of gamma and C with the best accuracy?

Thomas_Ott

@Kian you should try to search through the threads here, there's a bunch of examples. Or you could check out this example: https://www.neuralmarkettrends.com/predicting-historical-volatility-for-the-sp500/

earmijo

I also recommend reading: Practical Guide to Support Vector Classification

In short, they recommend a grid search for gamma = {2^-15, 2^-13, 2^-11, ... 2^3} and C = {2^-5, 2^-3, ..., 2^15}.