[Spambayes] very impressed....
tim at zope.com
Tue Jun 17 21:11:04 EDT 2003
[Eric C. Vogel]
> For me I find the sorting to by so so, still get a lot of spam in the
> inbox. Seems 90% still do not get caught.
That's terrible performance. You should post more about which version of
the software you're using, how you trained it, etc -- something is screwed
> I am amazed with your results. I have been running the application
> for over a month now and it virtually captures every spam coming to
> me. Something like 400 per day. I followed the install and
> teaching/learning instructions exactly after moving a couple of day's
> worth of SPAM to my SPAM folder then ran the learn feature. Now I
> just sit back and watch the junk count up in my SPAM folder.
> I get about 3 or 4 a day which are unsure and upon reviewing them I
> can understand the rating of Unsure since they could go either way.
OTOH, that's suspiciously good performance <wink>. An Unsure rate of under
1% isn't unheard of, but has rarely been seen in personal email. I got
rates that low in the early days, testing on a corpus of tech mailing-list
msgs. No "commercial" email of any kind counted as ham on that list, so the
job was much easier than sorting out my own email (which includes lots of
ham from, e.g., online merchants and financial institutions, much of which
is trying to sell me something). Whatever you're doing, keep doing it!
> Absolutely outstanding application... And I have tested a number of
Believe it or not, it's more the result of sweat than of magic. The March
issue of "Linux Journal" has an article by Gary Robinson about the
mathematical underpinnings of this project, and Richie Hindle has another
article there about the deployment aspects. It's easy when you know the
More information about the Spambayes