|-----Original Message----- | |I can't say I've looked at very many of the 100,000 hams. I |have a quarantine area where I can skim through the spam and |borderline stuff, but I don't keep a copy of the ham. |However, to be learned as ham, the Nigerian messages would |have to score below 0.5, and I don't think that's likely. Of |course, there could be other messages that have some of the |same tokens as Nigerian messages and that are being scored as |ham. But they might actually BE ham. | |> Lower your BAYES_00 score? (Towards zero, that is) | |That's what I'm doing unless I can find something better. | If you can't verify that hams are in fact hams you should probably turn autolearn off and feed sa-learn with some ham messages. This is very important at the beginning of training when the bayes dosent know ham from spam but after time it gets more and more accurate and auto-learn is pretty reliable then.
You can try to feed the nigerian spams into sa-learn and see if that corrects it, but it might be quicker to dump the bayes db and start again with autolearn turned off to start with.
