|
|
|
|
|
|
|
|
Applies To: Models 1000, 2000, 3000
STA (Statistical Token Analysis) uses several sources for "learning:"
If a message is flagged as spam by RBL or DCC, STA treats the message as spam.
STA also learns on your pattern based message filters. If you have a rule with an action of spam, STA will treat the message as spam. If you have it as trust or accept, it will treat the message as valid mail.
If you are running Update 2, and using Webmail along with RBL, DCC and STA, if Webmail flags a message as spam, STA will learn and treat this message as spam.
STA learns "good mail" by learning from trusted sources, such as your internal mail server.
