> Hello again! > > I've finally fixed some of my problems with DSpam, but was wondering if > anyone could help with another query. > > Basically, I have a setup which is hosting mail for lots of user (around > 50 domains and a total of 2000 email addresses are seen by dspam). Not > everyone is going to train the filter as simple put: they won't. So a > core group of us will be training it and then I plan to use dspam_merge > to merge this data into a globaluser which is part of a merged group. > This seems to be the right way to go, but then, everyone's own database > is being trained from mail it sees. So for a user who is not training > DSpam, it might flag lots of emails as not spam, thus eventually provide > poor spam catching. > > For instance if a user gets 100 emails through with the word 'casino' in > it, which are all spam but dspam doesn't know this yet, then dspam will > increase the innocent count on 'casino', thus eventually providing poor > accuracy down the line. > > Is it that I should be using TOE for the people who arn't part of the > core group of trainers? Is that the correct answer? > > Just trying to get my head around DSpam - I think it's great, and I just > want to get it working as best as possible. > > Regards, > Matt > > > > > From - Wed It sounds like you would want the "notrain" mode >From the README:
notrain: No training. Do not train the user's data, and do not keep totals. This should only be used in cases where you want to process mail for a particular user (based on a group, for example), but don't want the user to accumulate any learning data. David !DSPAM:1011,4952454c150924352620472!