A Probabilistic Theory of Pattern Recognition by Luc Devroye, László Györfi, and Gábor Lugosi is a graduate‑level monograph on the mathematical foundations of statistical pattern classification, with an emphasis on probabilistic analysis of learning rules.