m Borland
Cc: [EMAIL PROTECTED]
Subject: Re: [SAtalk] Training based on previously detected spam
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hello Graham,
Thursday, December 11, 2003, 7:21:52 AM, you wrote:
GB> I'm about to start training sa with spam and ham. Is it a good or
bad id
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hello Graham,
Thursday, December 11, 2003, 7:21:52 AM, you wrote:
GB> I'm about to start training sa with spam and ham. Is it a good or
bad idea
GB> to use spam which has already been detected by sa's non-trained
tests, or is
GB> that a waste of time
At 10:21 AM 12/11/2003, Graham Borland wrote:
I'm about to start training sa with spam and ham. Is it a good or bad idea
to use spam which has already been detected by sa's non-trained tests, or is
that a waste of time? Should I only train it with spam which is currently
managing to slip through?
W
At 11:45 PM 12/9/03 -0800, Justin Mason wrote:
Oh yeah, forgot to mention I finally got around to migrating all the FAQ
stuff onto the Wiki ;)
Heh, yeah, I caused me to go "Where the heck is that FAQ link???!!!" for
about 5 seconds before I saw the wiki one..
Ok, my real impressions were a little
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Matt Kettler writes:
>and then the Bayes FAQ in the Wiki:
>http://wiki.spamassassin.org/w/BayesFaq
>in particular:
>http://wiki.spamassassin.org/w/SiteWideBayesFeedback
>and:
>http://wiki.spamassassin.org/w/UsingAnAccountForLearning
Oh yeah, forgot t
At 02:01 AM 12/10/03 -0500, kittonian wrote:
Since the clients all download the mail and it becomes stored on the
exchange server for the user's inbox, how exactly do I train SA to stop
marking certain items? Our users are all over the place so if there's
something I can setup where I can have
On Tue, Nov 11, 2003 at 09:39:42AM +, Martyn Drake wrote:
> I was wondering how I can use sa-learn to train a system-wide Bayesian
> database? The way I've set-up SA at the moment is that each user has a
> .forward file which runs "|/usr/bin/procmail -f-" and that in turn reads
> the .pro
Hi,
On Fri, 7 Nov 2003, Evan Platt wrote:
> --On Friday, November 07, 2003 12:06 PM -0800 Jeff Lasman
> <[EMAIL PROTECTED]> wrote:
>
> > What's likely to happen if I run these 563 spams without running any
> > hams?
>
> I thought I recall the help file stating that running HAMS without SPAM (or
>
--On Friday, November 07, 2003 12:06 PM -0800 Jeff Lasman
<[EMAIL PROTECTED]> wrote:
> Since my original training a few weeks ago I've gotten 563 spams in my
> "spamuncaught" folder for further bayes training.
>
> What I don't have, is 500 or so ham emails that haven't been run through
> befo
Lucas Albers wrote:
Their is another potential method you could use.
Use the isbg.py imap python script to access the folders via imap and
learn the spam.
http://www.rogerbinns.com/isbg/
You could also use winbind to use the correcut NT usernames for the imap
connection from the linux box.
http://
Their is another potential method you could use.
Use the isbg.py imap python script to access the folders via imap and
learn the spam.
http://www.rogerbinns.com/isbg/
You could also use winbind to use the correcut NT usernames for the imap
connection from the linux box.
http://nic-ks.greatplains.n
I have kind of the same problem... I want my HAM to be sa-learned into
sa but the messages are on a store other then the Linux box, and
forwarding them will seem as if they are coming from an internal legit user.
What format of file does sa-learn expect and disect ?
-Maciek
PS Francisco, sorry t
On Wed, Jan 08, 2003 at 09:16:09PM +0100, Stefan Seiz wrote:
> My guess is that sa-learnspam will filter out and ignore the spamassassin
> specific header and subject prefix, but i am not sure about it (haven't
> found anything in the docs either).
From a quick look at the code, cleaning will happ
On Mon, Dec 09, 2002 at 11:22:20AM -0800, JP Kelly wrote:
> Is it possible to train SA 2.43 with known spam like sa-learn-spam does in
> SA 2.50?
2.43 doesn't have any "learning" rules. 2.50 has a bayes-ish system
which is what sa-learn-spam trains.
--
Randomly Generated Tagline:
How do I type
14 matches
Mail list logo
