Unsupervised learning algorithms

In an unsupervised scenario, as it's easy to imagine, there is no hidden teacher, hence the main goals cannot be related to minimizing the prediction error with respect to the ground truth. Indeed, the same concept of ground truth has a slightly different meaning in this context. In fact, when working with classifiers, we want to have a null error for the training samples (meaning that other classes than the true ones are never accepted as correct).

Conversely, in an unsupervised problem, we want the model to learn some pieces of information without any formal indication. This condition implies that the only elements that can be learned are the ones contained in the samples themselves. Therefore, an unsupervised algorithm is usually aimed at discovering the similarities and patterns among samples or reproducing an input distribution given a set of vectors drawn from it. Let's now analyze some of the most common categories of unsupervised models.