
guenther at rudersport
Oct 5, 2009, 11:17 AM
Post #6 of 9
(741 views)
Permalink
|
On Mon, 2009-10-05 at 11:01 -0700, Jefferson Davis wrote: > Thanks for the tips and low-grade knuck-wrap. Investigating - > installed 20_sought, tweaked local.cf back to 5.0 per list > recommendation. > > Appears that perhaps bayes_db is jacked up. re-training. All good. :) Just a minor nit, in case it isn't just different terminology. Installed sounds like a one-time operation -- the Sought rule-set needs to be updated using sa-update frequently, preferably more than once a day. Also note that when using sa-update, you need to run it for the stock rules, too. At the very least, once. A plain 'sa-update' will populate the stock channel. After training Bayes, it might be worth to keep an eye on it. In particular, that your users are not just deleting low-scoring spam, but instead move it somewhere after human review for training. -- 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; }}}
|