On Tue, Oct 26, 2004 at 06:19:40AM -0500, drunken-wallaby wrote: > hi there. is it somehow possible to estimate when evolution + sa start > filtering spam mails well? i've trained sa with at least 200 spam > mails, As I remember, it needs 200 spam *and* 200 non-spam before the Bayesian learner activates. -- Colin Watson [cjwatson at canonical.com]