To show the power of kernel SVMs, we're going to solve two problems. The first one is simpler but purely non-linear and the dataset is generated through the make_circles() built-in function:
from sklearn.datasets import make_circles nb_samples = 500 X, Y = make_circles(n_samples=nb_samples, noise=0.1)
A plot of this dataset is shown in the following graph:
Test dataset for an RBF SVM
As it's possible to see, a linear classifier can never separate the two sets and every approximation will contain on average 50% misclassifications. A logistic regression example is shown here:
from sklearn.linear_model import LogisticRegression ...
