"Re-importing" Mail Messages

For all users, who don't speak German!

Moderator: Forum-Team

"Re-importing" Mail Messages

Beitragvon Mwewe » 8. Okt 2004, 18:59

It's really a shame there's no way to re-analyze old messages. The technique of re-importing them means you end up with all your old messages duplicated in your mail client.

You could, of course, create another identity and receive all your messages into that identity, then you can delete it's messages and the identity itself after training.

But it leaves unnecessary additional tracks on your system that may be confusing at a later date.

When spam sneaks through the Learning Filter it no longer appears in the Recycle Bin by design. When you then check the bin and see only spam, you probably delete it and then go to the Learning Area.

After doing this a thousand times like I have, you can easily overlook a spam message incorrectly marked as non-spam and prematurely hit the training button by mistake. You've then taught the filter to consider a spam message as non-spam.

Do this enough times and you end up with enough errors in your word list to want to do something about it. Your choice of what to do is meager and depressing: 1. "re-import" all your old messages and re-train, 2. delete the wordlists and start over, or 3. live with it.

If you could edit the word filters it would make the Learning Filter more effective. If you could command the filter to analyze a folder of old mail messages without having to go through your mail client it would make re-training easy.

I've resorted to another solution. I put the training filter at the bottom of my priority chain. It almost never sees spam because the previous filters catch 99.9% of it. It's a small step now to simply disable the learning filter and I'll never miss it.

But it's a shame to do that because the Bayesian filter is one of the selling points for Spami.
Beiträge: 16
Registriert: 13. Mai 2004, 18:58

Zurück zu English Forum

Wer ist online?

Mitglieder in diesem Forum: Google [Bot] und 7 Gäste