You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Chris <cp...@embarqmail.com> on 2014/06/10 04:40:15 UTC

auto-learn

Since having to wipe my bayes db I've thought about going back to having
'auto-learn' setup for awhile. It's been so long since I did this I have
a fairly dumb question. Do I need the two below lines to be set and if
so is this the correct setting? Anything here about a score of 5 is
considered spam.

# bayes_auto_learn_threshold_nonspam 0.1
# bayes_auto_learn_threshold_spam 12.0

Thanks
Chris

-- 
Chris
KeyID 0xE372A7DA98E6705C
31.11°N 97.89°W (Elev. 1092 ft)
21:38:18 up 7 days, 6:08, 1 user, load average: 0.53, 0.45, 0.34
Mandriva Linux 2010.2, kernel 2.6.33.7-desktop586-2mnb


Re: auto-learn

Posted by Chris <cp...@embarqmail.com>.
On Tue, 2014-06-10 at 05:13 +0200, Karsten Bräckelmann wrote:
> On Mon, 2014-06-09 at 21:40 -0500, Chris wrote:
> > Since having to wipe my bayes db I've thought about going back to having
> > 'auto-learn' setup for awhile. It's been so long since I did this I have
> > a fairly dumb question. Do I need the two below lines to be set and if
> > so is this the correct setting? Anything here about a score of 5 is
> > considered spam.
> > 
> > # bayes_auto_learn_threshold_nonspam 0.1
> > # bayes_auto_learn_threshold_spam 12.0
> 
> Answering the direct questions first: Yes, that is correct syntax. No,
> you don't need them (commented out), they are default.
> 
> An auto-learning setup generally isn't a bad idea, and actually default.
> Depending on your amount of messages, you might want to have a look at
> the recent train-on-error option.
> 
> If (since) there was any need to wipe your old Bayes DB and start fresh,
> I seriously recommend continued manual training. And in any case, always
> (manually) training spam with low-ish Bayes probability. Likewise for
> ham that doesn't already have a very low Bayes probability.
> 
> In non-high-volume environments, there's hardly any down-side on
> training the extremes, too. Learning hand-confirmed non-extremes is
> always worth it.
> 
> 
> [1] http://spamassassin.apache.org/doc/Mail_SpamAssassin_Plugin_AutoLearnThreshold.html
> 
Thanks very much Karsten for the quick reply.

-- 
Chris
KeyID 0xE372A7DA98E6705C
31.11°N 97.89°W (Elev. 1092 ft)
22:18:08 up 7 days, 6:48, 1 user, load average: 0.56, 0.49, 0.61
Mandriva Linux 2010.2, kernel 2.6.33.7-desktop586-2mnb


Re: auto-learn

Posted by Karsten Bräckelmann <gu...@rudersport.de>.
On Mon, 2014-06-09 at 21:40 -0500, Chris wrote:
> Since having to wipe my bayes db I've thought about going back to having
> 'auto-learn' setup for awhile. It's been so long since I did this I have
> a fairly dumb question. Do I need the two below lines to be set and if
> so is this the correct setting? Anything here about a score of 5 is
> considered spam.
> 
> # bayes_auto_learn_threshold_nonspam 0.1
> # bayes_auto_learn_threshold_spam 12.0

Answering the direct questions first: Yes, that is correct syntax. No,
you don't need them (commented out), they are default.

An auto-learning setup generally isn't a bad idea, and actually default.
Depending on your amount of messages, you might want to have a look at
the recent train-on-error option.

If (since) there was any need to wipe your old Bayes DB and start fresh,
I seriously recommend continued manual training. And in any case, always
(manually) training spam with low-ish Bayes probability. Likewise for
ham that doesn't already have a very low Bayes probability.

In non-high-volume environments, there's hardly any down-side on
training the extremes, too. Learning hand-confirmed non-extremes is
always worth it.


[1] http://spamassassin.apache.org/doc/Mail_SpamAssassin_Plugin_AutoLearnThreshold.html

-- 
char *t="\10pse\0r\0dtu\0.@ghno\x4e\xc8\x79\xf4\xab\x51\x8a\x10\xf4\xf4\xc4";
main(){ char h,m=h=*t++,*x=t+2*h,c,i,l=*x,s=0; for (i=0;i<l;i++){ i%8? c<<=1:
(c=*++x); c&128 && (s+=h); if (!(h>>=1)||!t[s+h]){ putchar(t[s]);h=m;s=0; }}}