Hi,
Does anyone have any evidence (empircal or anecdotal) as to how well
2.5x fares against 2.4x using Bayes sitewide with minimal training (no
more than SA does when autolearning)?
It's my understanding (BICBW) that most of the PHRASE rules disappeared
in 2.5x in favour of Bayes. Which is all w
> I think you're going to run into problems. I would suggest you think
> about adding three more of those boxes, maybe more.
As a reference, 4 880s with 2 CPUs and 4GB of RAM are currently handling
~ 2.2 million messages a day (in conjunction with MIMEDefang) without
breaking in to a serious swe
> Does this Microsoft patent cause trouble?
>
> Here's the link to the document.
>
> http://makeashorterlink.com/?V27615462
ifile, at
http://www.nongnu.org/ifile/faq.html
has been doing client side bayesian filtering since 1996. That should
be sufficient prior art.
N
--
11
> As you can see, almost the same proportion of mail is scored above 5.0
> (I would expect this) - but now only a fraction of said mail is being
> blocked as spam. To get the same level of blocking on that second day,
> under 2.42, I actually have to lower the threshold from 10 to
> 6.6 - not a s
A couple of things I've observed in the recent past that I thought I'd pass
on.
adversend.com seem to exist solely to send adverts. Which makes deciding
what
to do with the mail very easy :-)
And the spammers seem to be getting a little more creative with their
message text to try and avoid s
