What's the best way to process missed spams ?

Hi all,

I use spamprobe in 'train' mode. Following what's described in the manpage, I 
have a script which calls 'spamprobe spam' on all spams which were missed (I 
don't call 'remove', though, as I recall reading that it wasn't necessary - 
perhaps I'm wrong). In anycase, this script doesn't seem to help in anyway. 

If I look at the terms which are used to score a missed spam before and after 
I call 'spamprobe spam' on it, almost all the terms are from the headers, 
even though very spammy terms (at least I'd think so) are readable in the 
message subject and body.

For instance, here's the output of '-T score' on a missed spam :

GOOD 0.0001013 adfa24bf07250a84e116d550070ed9e1
    Spam Prob   Count    Good    Spam  Word
    0.0000010       1     102       0  libraries
    0.0000025       1      32       0  what do
    0.0000030       1      27       0  dark
    0.0000058       1      14       0  Hsubject_spread
    0.0000068       1      12       0  everywhere
    0.0000081       1      10       0  work and
    0.0000090       1       9       0  brutally
    0.0000101       1       8       0  they need
    0.0000101       1       8       0  what they
    0.0000116       1       7       0  the dark
    0.0000162       1       5       0  and libraries
    0.0000162       1       5       0  needless
    0.0000162       1       5       0  needless to
    0.0000162       1       5       0  to ride
    0.0000203       1       4       0  and tired
    0.0000203       1       4       0  ending
    0.0000203       1       4       0  sick and
    0.0000203       1       4       0  they want
    0.0000270       1       3       0  about work
    0.0000270       1       3       0  all they
    0.0000270       1       3       0  fucked
    0.0000270       1       3       0  is everywhere
    0.9999676       1       0       5  Hrecvx_dsl.telesp.net.br
    0.0108696       1      91       2  ride
    0.0136986       1     180       5  that s
    0.9760120       1      16    1302  U_php
    0.0447761       1      32       3  do the
COMMAND score FILE spam DIGEST adfa24bf07250a84e116d550070ed9e1 SUBJECT Horny 
college girls ... PROCESSED

Then I remove it (just to make sure), flag it as spam and rescore it :

glaurent@linux4:glaurent> spamprobe remove spam
glaurent@linux4:glaurent> spamprobe spam spam
glaurent@linux4:glaurent> spamprobe -v -T score spam
SPAM 0.9972383 adfa24bf07250a84e116d550070ed9e1
    Spam Prob   Count    Good    Spam  Word
    0.9999838       1       0      10  Hrecvx_dsl.telesp.net.br
    0.9999676       1       0       5  Hfrom_ozondticxas
    0.9999676       1       0       5  Hfrom_ozondticxas yahoo.com
    0.9999676       1       0       5  Hrecv_033b590f1a
    0.9999676       1       0       5  Hrecv_033b590f1a for
    0.9999676       1       0       5  Hrecv_id 033b590f1a
    0.9999676       1       0       5  Hrecvx_0005fs
    0.9999676       1       0       5  Hrecvx_0005fs-uv
    0.9999676       1       0       5  Hrecvx_1dumus
    0.9999676       1       0       5  Hrecvx_1dumus-0005fs-uv
    0.9999676       1       0       5  Hrecvx_1dumus-0005fs-uv for
    0.9999676       1       0       5  Hrecvx_201-1-55-182.dsl.telesp.net.br
    0.9999676       1       0       5  Hrecvx_201-1-55-182.dsl.telesp.net.br 
201.1.55.182
    0.9999676       1       0       5  Hrecvx_201.1.55.182
    0.9999676       1       0       5  Hrecvx_201.1.55.182 helo
    0.9999676       1       0       5  Hrecvx_from 
201-1-55-182.dsl.telesp.net.br
    0.9999676       1       0       5  Hrecvx_id 1dumus-0005fs-uv
    0.0239090       1     102       5  libraries
    0.9760872       1      16    1307  U_php
    0.0270108       1     180      10  that s
    0.0370150       1      91       7  ride
    0.0564282       1    1227      97  search
    0.0642486       1     393      54  find
    0.0667873       1     391      56  work
    0.0724223       1      32       5  what do
    0.9245472       1      23     564  U_mail
    0.0846979       1      27       5  dark
COMMAND score FILE spam DIGEST adfa24bf07250a84e116d550070ed9e1 SUBJECT Horny 
college girls ... PROCESSED

Help welcome :-)

-- 
Guillaume.
http://www.telegraph-road.org


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click