What is the role of Naïve Bayes in probabilistic ensemble stacking systems?

Naïve Bayes is often used as a base learner in stacking ensembles to provide fast probabilistic signals.

In stacking, multiple base models generate predictions that are fed into a meta-model. Naïve Bayes can serve as a lightweight probabilistic base learner, especially useful when features are sparse. Its outputs can complement nonlinear models by providing independent probabilistic structure.