Nov 23, 2020 
Support Center » Knowledgebase » I've been training the spam filter but I'm still receiving the same spam emails
 I've been training the spam filter but I'm still receiving the same spam emails
Solution
Problem
 
Users have been training Kerio MailServer by marking messages as spam, or moving messages to the Junk Email folder, however they continue to receive the same spam emails.
 
Solution
 
Kerio MailServer includes a distribution of SpamAssassin, which consists primarily of two components: Static tests and dynamic tests. The tests are comprised of spam signatures that are compared to each message. A cumulative scored is assigned to each message. Spam emails will receive a higher score. By default, Kerio MailServer will consider any email which receives a score of 5.0 or higher to be spam.
 
The static tests are updated regularly, and compiled into new releases of Kerio MailServer. More information regarding the specific Spamassassin tests is available on the Spamassassin website
 
The dynamic or Bayesian tests are adjusted over time based on user feedback. The Bayesian database is updated when a user marks a message as 'spam' or 'not spam', or moves a message from their Inbox to the Junk Email folder and vice versa.
 
It is important to understand how the Bayesian tests really work: It does not outright flag messages as spam if they contain a specific subject, or sender address. It is only collecting specific characteristics of the message.
 
A message can only be flagged one time. If the same message is flagged multiple times, it will not affect anything as the dynamic tests have already been trained by that message.
 
The Bayesian tests are not active until it has received enough information. This includes a minimum of 200 spams and 200 hams (false positives).
 
The Bayesian tests are not always adjusted by user input. If a message receives a negative score, it is perceived as almost certainly NOT spam. Therefore, the entire message will be reprocessed to adjust the Bayesian tests. The same is true for the inverse situation, where a message is almost certainly spam.

Any message with a score of 12.0 or greater will satisfy this condition.
 
Once the Bayesian tests are active, most messages will contain a 'bayes' score, viewable in the X-Spam-Status header.


Article Details
Article ID: 546
Created On: Apr 25, 2006 03:48 PM

 This answer was helpful  This answer was not helpful

 Back
 Login [Lost Password] 
Email:
Password:
Remember Me:
 
 Search
 Article Options
Home | Register | Submit a Ticket | Knowledgebase | News | Downloads
Language: