[Spambayes] lots of unsures, heavily biased towards spam

skip at pobox.com skip at pobox.com
Mon Feb 5 18:06:43 CET 2007


    Dave> And deleting a messgae from the training set will cause it to be
    Dave> "untrained" at the next training?

If you retrain from scratch (as my train-to-exhaustion script does), it's a
non-issue.  If you use some sort of incremental training mechanism then the
message won't be there to be tokenized (so it won't know what token counts
to decrement).

Skip


More information about the SpamBayes mailing list