> I am still getting reports back
> from my users that emails from their customers are getting bounced back
> to them and flagged as spam.

Tell your users to send the customers an email - with the default setup 
(auto-white-listing is 'ON') the customers will be whitelisted and if 
'BayesWL' is 'OFF' (default) mails from these customers will be never 
check by HMM and Bayesian.


>How do I remove all previous entries out of the spamdb in MySQL?

Your DAILY rebuildspamdb task will replace the complete SPAMdb table (if 
'ReplaceOldSpamdb' is set to 'ON', which is the default)

Connect to the MySQL server

mysql> use assp;
mysql>delete from spamdb;

will remove all data from the MySQL table.

>Do I put DoBayesian in score mode?

This is the best choice in every case. I've set it one point below 
'PenaltyMessageLow'. (one point below 'PenaltyMessageLimit' is more 
agressive - flaggs SPAM, if confident SPAM is detected)
With one addionaly scoring point, the mail is flagged SPAM or possibly 
blocked by the penaltybox

>Currently it's set to block.

Yes - this is the case as long as your problem was posted here and at the 
forum.
This makes no sense, if the detection rate is bad (... a lot of false 
positives ...) - IMHO this must not be explained, because it is super 
logicial.

>DoHMM is turned off.

This makes even more no sense, because HMM is the much better engine.
If I would have your problem, I would set DoHMM at least to monitor, 
simply to see if it is wokring better than Bayesian or not.

Use 'DoPrivatSpamdb'.
Set 'baysConf' to 0.001 - this helps to prevent bad detection on short 
mails
make sure 'correctedspam' and 'correctednotspam' are set to default values

Jay, this (your) problem is going on for several weeks now - this is not 
possible.

If ...

- 'DoBayesian' were set to score below  'PenaltyMessageLimit' 
and
- all SPAM and all false positives were reported constantly (check the 
'errors' folder, there must be hundreds of mails in both subfolders - if 
not, nothing was reported) 
and
- the maintenance values for the short time corpus (spam , notspam folder) 
were set to a lower value (14 or 21 days for example) but at least to the 
defaults
and
- a DAILY rebuildspamdb task were configured

the problem would be gone.


Thomas




Von:    Jay <[email protected]>
An:     [email protected]
Datum:  23.09.2015 16:20
Betreff:        Re: [Assp-user] ASSP spam database start over?



Anyone have any ideas? I really have to take care of this and it's been 
5 days now.

On 9/18/2015 3:59 PM, Jay wrote:
> So I have a question, what would be the best way to go about starting my
> spamdb database over again from scratch? We are using MySQL as the
> database back end. The reason for this is I am still wrestling with the
> issue of ASSP being too sensitive and I am still getting reports back
> from my users that emails from their customers are getting bounced back
> to them and flagged as spam. Analyzing the emails have turned up that
> the message is flagged as spam even though the message may only contain
> 1 or 2 sentences pertaining to Job related information. ASSP got better
> since I restored from a backup from last year, but it's still too
> sensitive and I really need to address this.
>
> I believe I uncovered the issue with the database being too sensitive,
> it's the way the users have been forwarding the messages to the spam
> filter. Some were not sending the emails as an attachment, the emails
> have been sent inline. Most of my customers use Outlook and by default
> Outlook is set to send forwarded emails inline and not as an attachment.
> So for all these years some if not all my users have been submitting
> emails to the spam filter wrong. I am guessing because of this, the
> database now is too sensitive and causing the issues of a lot of false
> positives.
>
> Do I put the spam filter in test mode? How do I remove all previous
> entries out of the spamdb in MySQL? Just delete the database and start
> over? Do I put  DoBayesian in score mode? Currently it's set to block.
> DoHMM is turned off. Our corpus confidence comes out at 1.000000 which I
> know is perfect.
>
> Thank you for any assistance in advance.
>
>
>
> 
------------------------------------------------------------------------------
> _______________________________________________
> Assp-user mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/assp-user
>
>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2015.0.6140 / Virus Database: 4419/10661 - Release Date: 
09/18/15
>
>




------------------------------------------------------------------------------
Monitor Your Dynamic Infrastructure at Any Scale With Datadog!
Get real-time metrics from all of your servers, apps and tools
in one place.
SourceForge users - Click here to start your Free Trial of Datadog now!
http://pubads.g.doubleclick.net/gampad/clk?id=241902991&iu=/4140
_______________________________________________
Assp-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/assp-user






DISCLAIMER:
*******************************************************
This email and any files transmitted with it may be confidential, legally 
privileged and protected in law and are intended solely for the use of the 

individual to whom it is addressed.
This email was multiple times scanned for viruses. There should be no 
known virus in this email!
*******************************************************

------------------------------------------------------------------------------
Monitor Your Dynamic Infrastructure at Any Scale With Datadog!
Get real-time metrics from all of your servers, apps and tools
in one place.
SourceForge users - Click here to start your Free Trial of Datadog now!
http://pubads.g.doubleclick.net/gampad/clk?id=241902991&iu=/4140
_______________________________________________
Assp-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/assp-user

Reply via email to