RE: [Assp-user] Strange "Bayesian Spam"

[Theo A. <tau...@lo...>]

It is generally not usefull to post the actual text of you message,
since everyone's installation of ASSP will produce a slightly different
analysis.

My installation of assp shows the following when I run the text you
emailed.



Good Words=20
  blines from 0.0002=20
  server on 0.0019=20
  12 57 0.0067=20
  with smtp 0.0158=20
  12 55 0.0342=20
  2005 12 0.0373=20
  size 2 0.0555=20
  21 2005 0.0662=20
  10 2005 0.0833=20
  fri oct 0.2077=20
  face arial 0.2263=20
  arial size 0.2355=20
  format quoted-printable 0.2504=20
  font face 0.2933=20

Bad Words
00000000 ssub 0.9804  =20
href 0.8635  =20
charset iso-8859-1 0.8511  =20
x-mozilla-status2 00000000 0.8479  =20
21 10 0.8070  =20
for href 0.7758  =20
href it 0.7675  =20
helo 0.6971  =20

Analysis totals: 0.0002 0.0019 0.0067 0.0067 0.0158 0.9804 0.0342 0.0373
0.0555 0.0662 0.0662 0.0833 0.0833 0.8635 0.8576 0.8576 0.8511 0.8511
0.8479 0.8479 0.8070 0.8070 0.2077 0.2077 0.7758 0.2263 0.7675 0.2355
0.2504 0.2933 0.6971=20

spam-prob =3D 0.00000

I have separated the bad words from the good words.  When you put the
email into the analyze web page in ASSP, look at what it says as to why
it is spam.  Realize that it might produce a different result now then
it did before due to any database rebuild that might have occurred.

Theo