In the previous subsection we saw that solving the clustering problem implies solving a non-convex optimization problem that easily becomes a large-scale problem when having real-life datasets. Thus, we rely on heuristic methods to obtain reasonable solutions. One of the most used methods in the literature for clustering is k-means. We will learn how to implement it and analyse some of its limitations.

After studying this subsection, you should be able to:

• Implement k-means clustering.