Re: Re[2]: [Declude.JunkMail] Anyone with an updated Global.cfg?

2005-02-23 Thread Scott Fisher
-Mad,
Will there be an MDLP page explaining some of the columns?
SQ= Spam Test Quality?
SI = Spam Test Result Important Count?
avgSD = Average Spam Test Dominance?
- Original Message - 
From: "Pete McNeil" <[EMAIL PROTECTED]>
To: "Darrell ([EMAIL PROTECTED])" 
Sent: Wednesday, February 23, 2005 11:54 AM
Subject: Re[2]: [Declude.JunkMail] Anyone with an updated Global.cfg?


Just adding to the end of the thread here...
The demo of SNF is meant more to help you get things working on your
system than to prove it can capture spam. The demo rulebase is behind
the registered version quite a bit -- Folks have already told you that
though :-)
For a look at BLs to try and weights you might take a peek at a pet
project we're running. MDLP (Modular Declud Log Processing) is a
utility we currently have in beta that measures test accuracy. It also
includes an AI component to automatically adjust the weights of the
tests on a Declude system for best performance (with a little human
help occasionally).
You can see the HTML output of MDLP at the following link:
http://www.sortmonster.com/MDLP/
The page is under construction (really just a placeholder right now).
Report 1 is based on a small amount of data. Report 2 is based on a
larger amount of data. This runs and auto-tunes the tests on our
NT/Declude test bed which processes mostly spam, but a few user
accounts also to keep things honest.
You might use these reports to help evaluate some of the tests that
are out there and perhaps to suggest where the weights should be.
Best,
_M
(below retained for context)
On Tuesday, February 22, 2005, 7:13:23 PM, Darrell wrote:
Dsic> Ben,
Dsic> FYI - The full (licensed) version of Sniffer is incredible.  It 
grabs 95%+
Dsic> of spam we get right off the bat.  I highly recommend Sniffer.  The 
trial
Dsic> version's rule base lags the rule base you get when you are a 
licensed
Dsic> customers.


Hi,
I've noticed in the last couple of weeks a huge upsurge in junk mail
Dsic> getting
through our system with lower weights (i.e., ending up in the InBox
Dsic> instead
of the spam folder or being deleted).  We don't do a lot of tweaking 
with
our configuration files, so we normally expect a certain small 
percentage

---
[This E-mail was scanned for viruses by Declude Virus 
(http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail".  The archives can be found
at http://www.mail-archive.com.
---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]
---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail".  The archives can be found
at http://www.mail-archive.com.


Re[2]: [Declude.JunkMail] Anyone with an updated Global.cfg?

2005-02-23 Thread Pete McNeil
Just adding to the end of the thread here...

The demo of SNF is meant more to help you get things working on your
system than to prove it can capture spam. The demo rulebase is behind
the registered version quite a bit -- Folks have already told you that
though :-)

For a look at BLs to try and weights you might take a peek at a pet
project we're running. MDLP (Modular Declud Log Processing) is a
utility we currently have in beta that measures test accuracy. It also
includes an AI component to automatically adjust the weights of the
tests on a Declude system for best performance (with a little human
help occasionally).

You can see the HTML output of MDLP at the following link:

http://www.sortmonster.com/MDLP/

The page is under construction (really just a placeholder right now).
Report 1 is based on a small amount of data. Report 2 is based on a
larger amount of data. This runs and auto-tunes the tests on our
NT/Declude test bed which processes mostly spam, but a few user
accounts also to keep things honest.

You might use these reports to help evaluate some of the tests that
are out there and perhaps to suggest where the weights should be.

Best,

_M

(below retained for context)

On Tuesday, February 22, 2005, 7:13:23 PM, Darrell wrote:

Dsic> Ben,

Dsic> FYI - The full (licensed) version of Sniffer is incredible.  It grabs 95%+
Dsic> of spam we get right off the bat.  I highly recommend Sniffer.  The trial
Dsic> version's rule base lags the rule base you get when you are a licensed
Dsic> customers.




>> Hi,
>>
>> I've noticed in the last couple of weeks a huge upsurge in junk mail
Dsic> getting
>> through our system with lower weights (i.e., ending up in the InBox
Dsic> instead
>> of the spam folder or being deleted).  We don't do a lot of tweaking with
>> our configuration files, so we normally expect a certain small percentage




---
[This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)]

---
This E-mail came from the Declude.JunkMail mailing list.  To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail".  The archives can be found
at http://www.mail-archive.com.