Compare different kernels

Whenever using an SVM, we must specify a kernel to use. All of these kernels carry with them a set of potential hyper-parameters that the user needs to choose values for. The most common candidates are the classic, or linear kernel, SVM, RBF kernel SVM, and polynomial kernel SVM. The most common way to choose a kernel and set of hyperparameters to use cross validation to select the best working. Consider our cross-validation (link; topic and all tutorials and exercises)