[Spambayes] Retraining

Fred Mertz fred at lucy.com
Sun Feb 22 22:14:21 EST 2004


>> Now that I think I understand a little more about 
>> how SpamBayes works, I would rather have discarded a few of 
>> the messages I trained on.
>
>Ah - this is a bit different, though.  This is 'untraining', rather than
>'retraining'.  There isn't any facility exposed by the ui for this at the
>moment, sorry.  Unless you write some Python code and do it that way, the
>only option is to delete the databases and start from scratch.

Oh, ok.  Before I start from scratch, then, maybe I should ask about spams
with hundreds of random or gibberish words.  Do these muck up the 
databases?  They seem to be classified ok, and the clues aren't any 
of these gibberish tokens, but I know I trained on a few of these and it seems 
like they would skew the statistics.  What's the recommendation on this
type of spam?




More information about the Spambayes mailing list