NB-naive bayes

We use Naïve Bayes in many cases anyway, and
it often works pretty well
– often the right classification, even when not the right
probability (see [Domingos&Pazzani, 1996])