Social network analysis for email classification

ABSTRACT

The availability of a large corpus of emails in organizations, such as the Enron dataset (used in this work), is the motivation for this work. The attempt is to see if one can predict the organizational structure of Enron by using data mining algorithms and methodologies on this email dataset. The primary approach in this attempt is the analysis of email flows within the organization. Our results show that significant information about an organization's structure can be obtained even if the body (content) of emails is neglected. Enough relevant data is extracted about the 'email' social network using simple email flow analysis and associated statistics gaining an over all picture of the organizational structure. The longer term objective of this work is to show that readily available information can be used to determine relevant metrics by which one can reconstruct and verify the approximate social hierarchies within an organization or company.

REFERENCES

• 1: J. Diesner and K. M. Carley, "Exploration of Communication Networks from the Enron Email Corpus" Carnegie Mellon University.
• 2: B. Klimt and Y. Yang, The Enron Corpus: A New Dataset for Email Classification Research, ECML 2004.
• 3: W. W. Cohen, CALD, CMU. Retrieved October 5, 2004, from http://www-2.cs.cmu.edu/~enron/
• 4: J. Shetty, and J. Adibi, The Enron Dataset Database Schema and Brief Statistical Report. Retr. Nov. 4 2004, http://www.isi.edu/~adibi/Enron/Enron_Dataset_Rep.
• 5: http://www.jstor.org/view/00359254/di993342/99p04867/0?frame=noframe&userID=[e-mail]/01cce44061005014569b&dpi=3&config=jstor
• 6: S. Martin, A. Sewani, B. Nelson, K. Chen, A. D. Joseph. Analyzing Behavioral Features for Email Classification, Proceedings of the IEEE Second Conference on Email and Anti-Spam (CEAS 2005), July, 2005.
• 7: L. Yu*, K. R. Al-asmari, and S. Ramaswamy. The Dynamics of Open-Source Project Developer Network, article submitted for publication.
• 8: R. Popping, (2000). Computer-assisted Text Analysis. Thousand Oaks, CA: Sage Publications.
• 9: http://home.dei.polimi.it/matteucc/Clustering/tutorial_html/hierarchical.html
• 10: http://www.resample.com/xlminer/help/HClst/HClst_intro.htm

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.4 INFORMATION SYSTEMS APPLICATIONS

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
  I.7 DOCUMENT AND TEXT PROCESSING

Keywords:
business organization structures, data mining, mining e-mail archives