On Thu, 11 Aug 2011 15:35:02 +0200
Thierry de Coulon <tdecoulon@...> wrote:
Hello Thierry,
> Besides being slower, I also feel that SA tends to produce more "false
> positive" than Bogofilter - but I could not prove it. I can live with
It's been years since I stopped using SA, and can't remember the details
of its accuracy.
I rarely see false positives with bogofilter. Mostly, things land in
the "unsure" pile. Those messages are more often than not spam,
though. The messages that almost always fail to be detected as spam are
those with blank subject header and a link to a web site as the only
text in the body.
{snip}
> I do this from time to time, it wipes away old keywords that may no
> more be usefull (spam techniques change), and usually after a few days
Good point; My wordlist.db is probably a bit out of date. I'll start
re-training, I think.
--
Regards _
/ ) "The blindingly obvious is
/ _)rad never immediately apparent"
The man in a tracksuit attacks me
I Predict A Riot - Kaiser Chiefs
Attachments: