Naive Bayes (H2O)
Naive Bayes is an elementary classification algorithm which makes prediction by supposing simple conditional independence assumptions on the features. Although very basic, it can lead to good classification (but usually leads to poor probability estimations).
A regularization parameter, which smoothes conditional probability estimations, in case little data is available.
The minimum standard deviation allowed in a class (conditionally to the chosen feature) to perform a probability estimation. Higher values lead to more regularized models.