Re: Fundamental Training

"Charles A. Hewson" <[email protected]>
Newsgroups gmane.mail.bogofilter.general
Message-ID <[email protected]>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


On Sun, 8 Nov 2009, Volkan YAZICI wrote:

> Hi,
>
> I'm sure this dummy question had been asked dozens of times but I
> couldn't figure out the problem on my own via googling --- I really am
> not sure if it is really a problem. Anyway, here it goes.
>
>  $ rm -rf ~/.bogofilter         # Clean previous bogofilter registrations.
>  $ bogofilter -s </tmp/spam.0   # Initialize bogofilter wordlist database.
>  $ bogofilter -v </tmp/spam.1   # Test without any prior knowledge.
>  X-Bogosity: Unsure, tests=bogofilter, spamicity=0.520000, version=1.1.7
>  $ bogofilter -s </tmp/spam.1   # Train bogofilter.
>  $ bogofilter -v </tmp/spam.1   # Test after training.
>  X-Bogosity: Unsure, tests=bogofilter, spamicity=0.520000, version=1.1.7
>
> I trained my bogofilter database with similar ~300 messages, but I still
> get spamicity=0.520000 for any message I used at the training. I must
> have been missing something obvious. Could anybody help me with this
> problem?
>
>
> Regards.
> _______________________________________________
> Bogofilter mailing list
> [email protected]
> http://www.bogofilter.org/mailman/listinfo/bogofilter
>

You also need to train with some non-spam. I believe the suggested ratio 
is 1:4. Try 75 to 100 ham messages and scores will change.

$ bogofilter -n </tmp/ham.0
$ bogofilter -n </tmp/ham.1 .....
$ bogofilter -v </tmp/spam.###

Charles
- -- 
Charles Hewson <[email protected]>
Seattle, WA. U.S.A.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFK9hXOcVOO9PoMVFcRAkKzAKCmBALKlM8K5fidpsaesynR8xDRKQCePFvy
uuxNL4SHL8c8OIjcagGEtMA=
=ZArX
-----END PGP SIGNATURE-----
_______________________________________________
Bogofilter mailing list
[email protected]
http://www.bogofilter.org/mailman/listinfo/bogofilter
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.