On Thu, 16 Aug 2012 12:18:44 -0400 Alex wrote: > Hi, > > >> What will probably end up happening is this: > >> (1) wipe your Bayes database > >> (2) turn off autolearn > >> (3) collect several hundred hams and spams for an initial training > >> corpus (4) train using that corpus > >> (5) evaluate results > >> > >> Depending on your mail volume, once Bayes is working well after > >> manual training, you may then want to reenable autolearn; I > >> personally suggest it only where the volume is high enough and/or > >> the character of mail is varied enough to prohibit manual > >> training. You might also want to adjust the autolearn thresholds. > > What effect do whitelist entries have on autolearning
None at all because they are marked as "userconf". > In other words, my whitelist_from_rcvd entries add -100 to the score, > which would be way beyond the -3 I have required for autolearn. Setting a threshold of -3 is a bad idea unless you are going to write a lot of local rules with negative scores. The OP would be much better off zeroing the scores of the the offending DNSWL rules.
