In Chapter 7, Machine Learning For Recommender Systems, we learned about supervised machine learning. We used various features in the data (such as the user's ratings) to perform classification tasks. In supervised machine learning, we act a bit like a teacher—we provide a multitude of examples to our algorithm, which, once it gets enough data (and so its training is complete), is able to make generalizations about new items and infer their category or class.
But not all of the data lends itself to these kinds of tasks. Sometimes our data isn't labeled in any way. Imagine items as diverse as a website's traffic logs or the appointments made by customers at a dental clinic. These are just raw observations that aren't categorized in any way and don't contain any meaning. In such cases, data analysts employ unsupervised machine learning algorithms.
Unsupervised machine learning is used to discover hidden structures and patterns in otherwise unlabeled data. It is...