Given a set of attribute values, a probabilistic classifier is able to predict a distribution over a set of classes, rather than an exact class. This can be used as a degree of certainty; that is, how sure the classifier is about its prediction. The most basic classifier is Naive Bayes, which happens to be the optimal classifier if, and only if, the attributes are conditionally independent. Unfortunately, this is extremely rare in practice.

There is an enormous subfield denoted as probabilistic graphical models, comprising hundreds of algorithms for example, Bayesian networks, dynamic Bayesian networks, hidden Markov models, and conditional random fields that can handle not only specific relationships between attributes, but also temporal dependencies. Kiran R Karkera wrote an excellent introductory book on this topic, Building Probabilistic Graphical Models with Python, Packt Publishing (2014), while Koller and Friedman published a comprehensive theory bible, Probabilistic Graphical Models, MIT Press (2009).