Hello Graham, Monday, June 21, 2004, 9:22:21 AM, you wrote:
GD> Hello Andre, GD> Saturday, June 19, 2004, 12:25:11 PM, you wrote: AW>> Hello Richard, AW>> On 19 Jun 2004 at 05:33:38 +0500 GMT [02:33 CEST] you wrote: RHS>>> I'm using BayesIT, and have no problems with it. It's catching about RHS>>> 95% of the spam, which I hope will increase eventually, but even then RHS>>> that's only one or two a day that get through. AW>> Mine is at 99.8% for last month and 99.94% for last week :) GD> I would be happy if I could catch 1 spam....... :-( GD> I have SpamAssassin running on the mail server, it throws certain Spam into an IMAP GD> account for manual proofing - I mark all this and flag as Junk (training GD> BayesIt) GD> Spam that reaches TB also gets marked as Junk (more BayesIt training) GD> The problem is that BayesIt never catches spam on it's own !! GD> I did not have this problem with BayesIt 0.4g, I have followed the advice GD> to delete all previous dictionaries, but BayesIt just sits there and does GD> nothing I installed a newer version of Bayesit! yesterday, going from 0.4gm to 0.5.5. I see that the only difference is that you need to remove the earlier bases and that the training function in the configuration dialog is gone. I marked all the mails that I have in my Junk folders (about 25.000) and clicked "Mark as Junk" from the Specials Menu. That's the new way to train, it even shows a small dialog (if you have enabled it) that says training. Would have been fun to have some kind of documentation for 0.5.5, I found some in russian but babelfish.alstavista.com did not help me that much, and I did not understand anything of it :). If you do not have enough SPAM I could send you some :) for training. /Fredrik ________________________________________________ Current version is 2.11.02 | 'Using TBUDL' information: http://www.silverstones.com/thebat/TBUDLInfo.html
