The k-means algorithm

In this section, you will learn about how the k-means algorithm works under the hood, in order to cluster data into groups that make logical sense.

Let's consider a set of points, as illustrated in the following diagram:

A random set of points

Assignment of centroids

The first step that the algorithm takes is to assign a set of random centroids. Assuming that we want to find two distinct clusters or groups, the algorithm can assign two centroids, as shown in the following diagram:

Centroids, represented by stars

In the preceding diagram, the stars represent the centroids of the algorithm. Note that in this case, the clusters' centers perfectly fit the two distinct groups. This is the most...