bogofilter satisfaction report

[Lennart Sorensen]

On Fri, Apr 29, 2005 at 10:15:38AM -0400, JoeHill wrote:
> On Fri, 29 Apr 2005 08:55:46 -0400
> Lennart Sorensen disseminated the following:
> 
> > > Not givin' up yet!
> > 
> > I installed bogofilter 2 days ago, and I fed it my mailing lists and
> > inbox and spam folder like this
> > 
> > cat mailinglistmbox | bogofilter -M -n
> > cat spamfolder | bogofilter -M -s
> > 
> > After which it has been rather impresively accurate (I removed
> > spamassasin from my procmailrc when I put in bogofilter and so far so
> > good).
> > 
> > I think I probably fed it 10000 spam and 3000 non spam.
> 
> I tried doing it over again, with the -M switch, which I had not done before.
> It's still killing everything.
> 
> I checked the config file, and uncommented these:
> 
> ham_cutoff = 0.45
> spam_cutoff= 0.99
> 
> but still the same situation.
> 
> What the heck am I doing wrong here??

Remember to pass -n for nonspam training and -s for spam training.

You can then try and run bogofilter -T -M < nonspammbox to see what rating
(and score) it assigns to messages in that mailbox.  -TT gives just the
number in an easier to parse format.

Most of my normal mail is running in the range 0 to 0.04 with the
occational announcement list mail showing up at around 0.45.  The spam
seems to run around 0.997 to 1.

How big is the .bogofilter/wordlist.db?

Lennart Sorensen
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml