Kerio Tech Version 6.5.2 - Manual (Page 43)

1.5 Email
43
Figure 1.37 Dictionary
1.5.7 Spam filter
Spam messages are mostly represented by advertisements. Spam are usually sent in bulk and
the recipient addresses are obtained by illegal means (e.g. by tapping the network communi-
cation). In their wilder modifications, spams may also distribute and seed dangerous viruses.
Kerio MailServer contains a sophisticated antispam filter. The filter is designed so that it
can recognize specific attributes of spam from the message contents (spam messages usually
contain specific keywords that can be recognized and counted by the filter). Each message is
assigned a numeric score; if this value exceeds the limit set in Kerio MailServer, the message
is marked as spam.
Self-learning antispam filter
The built-in antispam filter may let some spam messages through from time to time by mistake
(their score is low) and mark some regular messages as spam (their score is too high). For the
reasons above, the antispam filter allows modification to the database used for recognition
of spam messages. This method, however, requires user input. Users should reassign the
incorrectly evaluated messages to correct types (spam / non-spam) so that the filter learns to
recognize them in the future.
Kerio WebMail uses the Spam and Not spam buttons located in the toolbar to mark the mes-
sages correctly. Highlight the incorrectly marked message and click one of the buttons. The
filter receives information regarding update of the database. This decreases chance for incor-
rect deliveries.