My guess is that if it has a high bayes score already, you don't need to re-learn it - the Bayesian engine already /knows/ about the interesting tokens contained therein.
However, I admit to sa-learning my lower-scoring spam that may have a BAYES_n where n is less than 90. If it has a BAYES_99 I don't need to sa-learn it. -tom > -----Original Message----- > From: Jon Gabrielson [mailto:[EMAIL PROTECTED] > Sent: Friday, August 29, 2003 1:03 PM > To: Tom Meunier; [EMAIL PROTECTED]; > [EMAIL PROTECTED] > Subject: Re: [SAtalk] Changing Bayes scoring > > > Is there a way to change this behavior? > It seems to me that a high bayes score also shows that it is spam > and it might be possible to grab a few new tokens from the spam > which you otherwise wouldn't get. > > Jon. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
