In this chapter, we've discussed an important unsupervised learning task: clustering. The simplest clustering algorithm is k-means. It doesn't provide stable results and is computationally complex, but this can be improved using k-means++. The algorithm can be applied to any data for which Euclidean distance is a meaningful measure, but the best area to apply it is a signal quantization. For instance, we've used it for image segmentation. Many more clustering algorithms exist for different types of tasks.
In the next chapter, we're going to explore unsupervised learning more deeply. Specifically, we're going to talk about algorithms for finding association rules in data: association learning.