Machine Learning With Go

Overview of k-means clustering

Let's say that we have a bunch of data points defined by two variables, x₁ and x₂. These data points naturally exhibit some grouping into clusters, as shown in the following figure:

To automatically cluster these points using k-means, we would first need to choose how many clusters will result from the clustering. This is the parameter k, which gives k-means its name. In this case, let's use k = 3.

We would then randomly choose the x₁ and x₂ locations of k centroids. These random centroids will serve as our starting point for the algorithm. Such random centroids are shown in the following figure via Xs:

To optimize ...