PCA is good at reducing the number of dimensions, but it works in a linear manner. If the data is not organized in a linear fashion, PCA fails to do the required job. This is where Kernel PCA comes into the picture. You can learn more about it at http://www.ics.uci.edu/~welling/classnotes/papers_class/Kernel-PCA.pdf. Let's see how to perform Kernel PCA on the input data and compare it to how PCA performs on the same data.
import numpy as np import matplotlib.pyplot as plt from sklearn.decomposition import PCA, KernelPCA from sklearn.datasets import make_circles