One of the main ideas behind collaborative filtering is that if user A has the same opinion of a product as user B, then A is also more likely to have the same opinion as B on another product than that of a randomly chosen user.
Clustering is one of the most popular techniques used in collaborative-filtering algorithms. It is a type of unsupervised learning that groups data points into different classes in such a way that data points belonging to a particular class are more similar to each other than data points belonging to different classes:
For example, imagine that all our users were plotted on a two-dimensional Cartesian plane, as shown in the preceding graph. The job of a clustering algorithm is to assign classes to every point on this plane. Just like the similarity measures, there is no one clustering algorithm to rule them all. Each algorithm has its specific...