| Bayesian filter threshold |
 Posted by
Jason Gryder
on Wednesday, December 31, 2008 at 12:48:47 PM (EST)
Is there a way to get the Bayesian filter threshold % below 90%? I change it from the default 90% and it returns to 90 after I hit save. Does there have to be a certain # of learned spam before you can adjust it? I have trained it for quite a while now. Thanks.
|
 Posted by
Robert Smith
on Friday, January 23, 2009 at 2:29:17 PM (EST)
It does the same thing for me, I think it's the lowest it can go. The more you train it, the better it will be at identifying spam.
|
Posted by
Barry Kucher
on Saturday, February 28, 2009 at 12:58:52 AM (EST)
I am surprised you can even get away with 90%. I have to keep it at 99% to make sure most emails get through. We get emails from a couple of banks, and if I put it even at 98%, these emails are marked as spam. Even after I add it to our whitelist, they get marked as spam if I put it below 99%.
Even at 99%, about 50% of the email we receive is flagged as spam. This is even after my mail host performs their spam filtering. Their spam filter removes about 40% of our incoming emails as spam.
At 90%, what percentage of your incoming emails are spam? I am curious. You can tell basically by going to the administrations main page and see how many emails were marked as spam and how many emails got thru.
We process about 2000 emails a day, and about 1000 are spam EVERY DAY!
Barry Kucher
www.411tech.org
|
|
