In this chapter, we talked about some unsupervised learning algorithms, including k-means, spherical clustering, and agglomerative hierarchical clustering. We saw that k-means is just a specific application of the more general expectation-maximization algorithm, and we discussed its potential limitations. Furthermore, we applied k-means to two specific applications, which were reducing the color palette of images and classifying handwritten digits.
In the next chapter, we will move back into the world of supervised learning and talk about some of the most powerful current machine learning algorithms: neural networks and deep learning.
