Abstract:
In a complex Email system with plenty of messages, each user can receive dozens of
messages every day. How to recognize which messages should be read first and how to
find important or prominent people in the system without knowing the messages’ content.
This study focuses on Email transactions that show relationships among entities via
messages’ transference in the system to discover important messages and important people
from email log. Important messages are messages that contain interesting information or
important contents and important people are those who have strong effects on the
community via sending and receiving several messages. Two data structures are introduced
for this task. The Email transaction multi-digraph is used to represent the email
transactions which are encoded in the mail log. Whereas, the messages’ flow is used by the
Scoring Model to calculate the scores for email messages based on types of messages
(original, forward, or reply). The results of the Scoring Model are used to determine
important messages and important people. Some experiments were carried out to verify the
methodology and the results showed that this model’s result is acceptable.
