What is the difference between generative and discriminative models in the context of Naïve Bayes?

Naïve Bayes is a generative model because it models joint probability P(X, Y), unlike discriminative models that model P(Y|X).

Generative models like Naïve Bayes learn how data is generated by estimating P(X|Y) and P(Y), enabling them to compute P(Y|X) via Bayes' theorem. Discriminative models like logistic regression directly estimate P(Y|X) without modeling data distribution. Generative models are typically more data-efficient but may be less accurate asymptotically.