info about spam messages
David Relson
relson at osagesoftware.com
Fri Jun 11 18:08:13 CEST 2004
On Fri, 11 Jun 2004 18:31:56 +0300
Tayfun ASKER wrote:
> Hi David,
> I don't use the "replace_nonascii_characters" option. I just train
> bogofilter with Turkish spam and nonspam. Bogofilter is quite
> successful in parsing Turkish words and catching Turkish spam.
>
>
> I hope this is the answer to your question.
>
> Regards,
>
> Tayfun
Yes. You've told me what I wanted to know.
I've heard that bogofilter's parsing of languages like chinese and
japanese is quite "broken", i.e. the way that the message is parsed into
tokens has little relation to what the words actually are. However a
google search for bogofilter gives a number of japanese sites, so it
must be doing something useful.
David
More information about the Bogofilter
mailing list