Peter Cheng wrote:
Hi Kyle Johnson,
Thanks for your quick reply.. really appreciate that!!
Below are my current "group" file configuration:-
[EMAIL PROTECTED] dspam]# cat group
GLOBALGROUP:classification:*globaluser
[EMAIL PROTECTED] dspam]#
[EMAIL PROTECTED] dspam]#
I should change it to this right ?
[EMAIL PROTECTED] dspam]# cat group
GLOBALGROUP:classification:*globaluser
[EMAIL PROTECTED]:merged:*
[EMAIL PROTECTED] dspam]#
[EMAIL PROTECTED] dspam]#
Besides, just want to know will this help ? Let say i change it as
below:-
[EMAIL PROTECTED] dspam]# cat group
GLOBALGROUP:classification:*globaluser
[EMAIL PROTECTED]:merged:*
[EMAIL PROTECTED]:shared:* <<--- what does it mean?
[EMAIL PROTECTED] dspam]#
[EMAIL PROTECTED] dspam]#
Kyle Johnson wrote:
Peter Cheng wrote:
Hi All,
Oi!
I really run out of idea how to configure DSPAM works this way.
Would be appreciate if anyone advised me how to configure my dspam
to works as following scenarios.
No problem!
I have configure the system to do the corpus database and it is
quite ok at the beginning.. but after a weeks it seem to be miss out
few spam.. those are using globaluser configuration.
Are you using already any type of group (classification, merged, etc)?
Because of that, i have done feed & train dspam on one email account
assume is "[EMAIL PROTECTED]". & now it is quite accurate.
Explain. When dspam makes an error, you shouldn't have to feed *or*
train dspam - you should need only to *re*train (dspam --source=error
--class=spam --user username --signature=signature).
Question:-
How do i configure everyone to refer to this database
"[EMAIL PROTECTED]" instate of their own database ?
For this, if you are not already, you will need to use a group.
There are quite a few types of groups (inoculation, merged, shared,
classification, etc), as laid out in the README, but you will most
likely want to use a merged group; to do so:
1. Create a file named group in dspam's home directory, which is most
likely /usr/local/var/dspam/.
2. Inside this file, one line: [EMAIL PROTECTED]:merged:*
This will tell dspam to, at run time, merge
[EMAIL PROTECTED]'s data with *.
Basically, I want to train one user "[EMAIL PROTECTED]" and
other users just to use that user's database.
See above. While this will provide out-of-the-box accuracy, users
will need to still retrain the errors that dspam makes for their
mail, or else their accuracy will fall.
To train, you should use dspam_train, with a corpus of spam AND ham,
in maildir format, such as dspam_train [EMAIL PROTECTED]
/path/to/spam/ /path/to/ham/
Thank you for the answer in advance..
Best Regards,
Peter Cheng
You are welcome, and I hope that this helps.
I am very inconsistent with these mailing lists, so if I do not
respond to a later question, do forgive me.
--Kyle
!DSPAM:45ece8c9232741800720552!
Best Regards,
Peter Cheng
[EMAIL PROTECTED] is only an example.
You want to create a new user that, and then run dspam_train on said
user. This will provide that user with some out-of-the-box accuracy.
You want to then add that user to a group, to provide that data to
others. I recommend a merged group, and if the created user's name (as
matches in the dspam_virtual_uids table) [EMAIL PROTECTED], the
group file would contain [EMAIL PROTECTED]:merged:*.
I am pretty sure that, with a shared group, all data is shared between
users. So, [EMAIL PROTECTED]:shared:* would be wrong - you
would have generalgroupnameforreference:shared:*. I think.
generalgroupnameforreference would then be created, and you can check
him out in the WebUI. I think - never used one (a shared group).
Hope this helps,
Kyle
