Drexel University Home Pagewww.drexel.edu DREXEL UNIVERSITY LIBRARIES HOMEPAGE >>

iDEA: Drexel E-repository and Archives > Drexel Academic Community > College of Information Science and Technology > Research Day Posters (IST) > Exploring a method of extracting universal features of phishing emails

Please use this identifier to cite or link to this item: http://hdl.handle.net/1860/1571

Title: Exploring a method of extracting universal features of phishing emails
Authors: Lee, Ki Jung
Song, Il-Yeol
Keywords: Phishing;Feature extraction;Dual process cognition;Elaboration likelihood model;Persuasion;Factor analysis;Logistic regression
Issue Date: 17-Apr-2007
Publisher: Drexel University. College of Information Science and Technology.
Series/Report no.: IST Research Day 2007 posters
Abstract: Current approaches of phishing filters depend on classifying emails based on obviously discernable features such as IP-based URLs or domain names. However, as those features can be easily extracted from a given phishing email, in the same sense, they can be easily manipulated by sophisticated phishers. Therefore, it is important that universal patterns of phishing messages should be identified to serve as a basis for novel phishing classification algorithm. In this paper, we argue that phishing is a kind of persuasion and explore feature extraction method based on persuasive communication perspective. Phishing message components, including message factors, source factors, and computer related factors, are investigated as message sender’s strategic message manipulation. On the other hand, message receiver’s cognitive components for information processing are discussed in terms of dual process of cognition. Our method consists of four major procedural steps. First, persuasive message components are identified through extensive literature review. Second, based on the identified persuasive message components, we conduct content analysis of email messages. Third, using factor analysis, persuasive components in phishing messages are classified for the validation of a dual process of cognition. From the pool of persuasive communication variables, we identify underlying dimensions to see whether central route information processing and peripheral route information processing are distinctly identified. Fourth, instances are classified by conducting logistic regression analysis based on the identified variables as a result of factor analysis in addition to known phishing factors identified by other studies. We, then, present a quantitative model that can represent persuasive information structure in phishing messages. This paper makes contribution to phishing classification research by presenting the idea of universal information structure in terms of persuasive communication theories.
URI: http://hdl.handle.net/1860/1571
Appears in Collections:Research Day Posters (IST)

Files in This Item:

File Description SizeFormat
2007021039.pdf351.44 kBAdobe PDFView/Open
View Statistics

Items in iDEA are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0! iDEA Software Copyright © 2002-2010  Duraspace - Feedback