Machine Learning Algorithms

Kernel PCA

We're going to discuss kernel methods in Chapter 7, Support Vector Machines, however, it's useful to mention the class KernelPCA, which performs a PCA with non-linearly separable data sets. Just to understand the logic of this approach (the mathematical formulation isn't very simple), it's useful to consider a projection of each sample into a particular space where the dataset becomes linearly separable. The components of this space correspond to the first, second, ... principal components and a kernel PCA algorithm, therefore, computes the projection of our samples onto each of them.

Let's consider a dataset made up of a circle with a blob inside:

from sklearn.datasets import make_circles>>> Xb, Yb = make_circles(n_samples=500, ...