Fortinet FortiMail 3.0 MR4 - Manual (Page 101)

set as control bayesian
FortiMail™ Secure Messaging Platform Version 3.0 MR4 CLI Reference
06-30004-0420-20080814
101
as control bayesian
Use these commands to set the names for Bayesian control accounts.
Syntax
set as control bayesian is-spam <name_str>
set as control bayesian is-not-spam <name_str>
set as control bayesian learn-is-spam <name_str>
set as control bayesian learn-is-not-spam <name_str>
set as control bayesian training-group <sender_str>
Example
An administrator wants to change two of the Bayesian control account names. He knows his users will
be better able to remember the addresses user to train the database with new messages if they
include the word ‘train’:
The learn-is-spam command becomes train-is-spam and the learn-is-not-spam command becomes
train-is-not-spam. To make these changes, enter these commands:
set as control bayesian learn-is-spam train-is-spam
set as control bayesian learn-is-not-spam train-is-not-spam
Keywords and Variables Description Default
is-spam FortiMail end users can send spam messages that were
incorrectly treated as non-spam to this account to inform
the Bayesian antispam check of its mistake. is-spam
is-not-spam FortiMail end users can send non-spam messages that
were incorrectly treated as spam to this account to inform
the Bayesian antispam check of its mistake. is-not-spam
learn-is-spam End users send known spam to this account to train the
FortiMail unit.
Based on the sender’s email address, the FortiMail unit
uses the information received to train the sender’s
Bayesian database. learn-is-spam
learn-is-not-spam End users send existing non-spam email to this account to
train the FortiMail unit.
Based on the sender’s email address, the FortiMail unit
uses the information received to train the sender’s
Bayesian database. learn-is-not-spam
training-group This account contains a system-wide spam database set up
by the administrator.
Using this account name as the “from” address, the
administrator sends confirmed spam to the “learn-is-spam”
user account and good email to the “learn-is -not -spam”
user account to do group Bayesian training.
If an individual user’s Bayesian database does not contain
sufficient information for spam scanning, it will use the data
received from the training group user account to scan
spam. default-grp
<name_str> This is the name for this account. Users send messages to the email address
composed of this name, followed by “@”, followed by the email domain.
<sender_str> This is the ‘from’ name used when sending mail to one of the other four accounts.
Mail can be sent to correct incorrectly categorized mail, or to train the Bayesian
database with new mail. Administrators send messages from the email address
composed of this name, followed by “@”, followed by the email domain.