David Jones wrote: > > James B. Byrne wrote: > > If you keep Bayes well trained (assuming you have enough ham to do so) > > Bayes poisoning is a myth. > > I'm not sure I agree with the "myth" statement. I just had to reset my Bayes > DB after years of it slowly drifting due to bad user input and such.
Years? How far back does your Bayes db store data? Mine shows: $ sa-learn --dump magic 0.000 0 3 0 non-token data: bayes db version 0.000 0 33982 0 non-token data: nspam 0.000 0 79295 0 non-token data: nham 0.000 0 125305 0 non-token data: ntokens 0.000 0 1397501146 0 non-token data: oldest atime 0.000 0 1400294391 0 non-token data: newest atime 0.000 0 0 0 non-token data: last journal sync atime 0.000 0 1400266810 0 non-token data: last expiry atime 0.000 0 2764800 0 non-token data: last expire atime delta 0.000 0 13546 0 non-token data: last expire reduction count $ date -R -d @1397501146 Mon, 14 Apr 2014 12:45:46 -0600 $ date -R -d @1400294391 Fri, 16 May 2014 20:39:51 -0600 My Bayes db only has the last month's of data in it. That is a completely stock configuration. I think the storage is actually by number of tokens not age though. It would be great if someone could explain that in better detail. > I added CRM114 and BOGOFILTER plugins as a "balance of power" to > Bayesian and it's working very well again. Both of those are good learning Bayes engines on their own. Bob