For a homework I used a support vector machines (classification) with a RBF kernel, k-fold cross validation=10, cost of constraints violation=100, and gamma=0.001.
It works great and my predictions are very good.
However, we didn't see how to formulate the model in mathematical terms (well we are not going to cover svm in the course at all), and I believe it's important.
I reviewed Hastie's and other books, but I still don't know how to formulate this model.
Can somebody point me in the right direction?