In v3.6.6 I have retrained the exact same message (Nigerian scam) about
20 times or more and it then doing a dspam --classify and it is still
showing up as "Innocent". I have over 80,000 messages in my corpus.
v3.4.9 never did this. I'm baffled. I would like to figure out a way
to see all the tokens and the scores for each it is using in the
calculation.
Charles
Abel Jeffcoat wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> All,
>
> I have one user that keeps getting the same Spam message over and
> over, even if it has been re-learned as Spam 4 times.
>
> I have attached it with headers. Any ideas?
>
> Abel Jeffcoat
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.3 (MingW32)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFEeNuv4MnK8L01kLMRAjuAAKC46gjjGDfwIMYNfigh71Lou0OVCACfZZ+V
> u2hQF0caRrzirSKz4C6As8M=
> =bKVz
> -----END PGP SIGNATURE-----
>
>
Received on Tue May 30 09:10:21 2006
This archive was generated by hypermail 2.1.8 : Wed May 31 2006 - 00:00:03 EDT