Open Access
Johansen, Lisa Yvette
Graduate Program:
Computer Science and Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
April 01, 2008
Committee Members:
  • Patrick Drew Mcdaniel, Thesis Advisor
  • domain reputation
  • social network
  • email
Email has become an integral and sometimes overwhelming part of users’ personal and professional lives. Due to the number of emails and their wide range of importance, managing this communication medium has become a intensely researched topic. In communication net- works, the identification of communities of interest (COIs) – groups of users that share a common bond – has proven highly applicable in automating various tasks. In this thesis, we measure the flow and frequency of user email toward the identification of COIs. Through this identification, we hope to enable the automation of some of the management tasks associated with email. We begin by analyzing a large corpus of university email in order to drive the development of algo- rithms for automatically determining COIs in email. We then validate the proposed algorithms by evaluating their ability to serve as an automated priority filter. Our analysis shows that the proposed algorithms correctly identify email as being sent from the human-identified COI with high accuracy. This indicates that a significant amount of information can be determined solely from the sender and receiver of an email. Identification of COIs in email communication can be highly applicable in a variety of email-related applications. In the second part of our research, we look at a possible application for COI: an automated reputation service for use with DKIM (DomainKeys Identified Mail). We describe a COI-based domain reputation service, analyze its ability to identify relationships between users and domains and compare its characteristics to current reputation services. We discuss the benefit of employing the identification of communities of interest in email.