How it works...

In this recipe, we have used an SVM classifier to find the best separating boundary between a dataset of points by solving a system of mathematical equations. To address a nonlinear problem, we used Kernel methods. Kernel methods are thus named for Kernel functions, which are used to operate in the feature space without calculating data coordinates in space, but rather by calculating the internal product between images of all copies of data in the function space. The calculation of the internal product is often computationally cheaper than the explicit calculation of the coordinates. This method is called the Kernel stratagem.