Lecture
This lecture covers the principles and steps of the Kernel K-means algorithm, including the derivation of the kernel version, the geometrical division of space using different kernels, and the advantages, sensitivity, hyperparameters, and limitations of K-means clustering. It also explores the concept of Kernel K-means to perform non-linear clustering with norm-2 in feature space, the objective function minimization, and the iterative procedure involved. The lecture concludes with an analysis of the terms in Kernel K-means, an exercise on partitioning space with different kernels, and the limitations and considerations when choosing the number of clusters.