Using CBR as Design Methodology for Developing Adaptable Decision Support Systems - Efficient Decision Support Systems: Practice and Challenges from Current to Future - page 139

Attribute Type Description

ID Numeric Unique message identifier.

From Alphanumeric Source mailbox.

Return Path Indicates the address that the message will be returned

to if one chose to reply.

Alphanumeric

Date dd-mm-yyyy Date in which the message was sent.

Language Alphanumeric Particular tongue of the message.

Attached Files Numeric Indicates the number of attached files.

Content Type Enumeration MIME type.

Relevant Terms Numeric Number of selected features to cluster the message.

Total Terms Numeric Number of features contained in the message.

Frequency-Term
Descriptor

Storing for each feature a measure of their frequency in
the message.
Class Enumeration Message category. Possible values are: spam , legitimate,
unknown .
Table 4. Structure of an instance representing an incoming e-mail in the SpamHunting
system

Array of feature-
frequency pairs

Fig. 10. Life cycle of the SpamHunting system and its integration with existing MTA and
MUA
Whenever SpamHunting receives a new e-mail, the system evolves through the four steps
depicted in the lower part of Figure 10 as shadowed rectangles. Initially the system