On Wed, 2014-05-14 at 11:38 +0100, Timothy Murphy wrote:
> [tim@grover ~]$ sa-learn --dump magic
>
> 0.000 0 3 0 non-token data: bayes db version
> 0.000 0 25758 0 non-token data: nspam
> 0.000 0 36434 0 non-token data: nham
> 0.000 0 144860 0 non-token data: ntokens
That sure is sufficient training (number of spam and ham messages, and
individual tokens learned).
The amount of ham might possibly skew results. But to see weather bayes
scores are biased towards hamminess, we'd need the X-Spam headers --
which I will not ask you a third time for.
> 0.000 0 1390675205 0 non-token data: oldest atime
> 0.000 0 1400062502 0 non-token data: newest atime
> 0.000 0 1400049904 0 non-token data: last journal sync atime
Last db access and journal sync times are recent, from today. Everything
looking fine.
If you still suspect Bayes to not work properly, you'll have to provide
more details.
--
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; }}}
