Link-based event detection in email communication networks

ABSTRACT

People's email communications can be modeled as graphs with vertices representing email accounts and edges representing email communications. Email communication data usually comes in as continuous data stream. Event detection aims to identify abnormal email communications that serve as analogs of real-world events imposed upon the data stream. The goal is to understand the communications behaviors of the subjects. The contents of emails are often not available or protected by privacy, which makes linkage information the only resource we can rely on. We propose a link-based event detection method that clusters vertices with similar communication patterns together and then, considers deviations from each vertex's individual profile, as well as its cluster profile. Experiments show that this method performs well on both Enron and our own email datasets.

REFERENCES

• 1: Scan statistics on enron graphs. http://cis.jhu.edu/parky/Enron/enron.html.
• 2: J. D. Banfield and A. E. Raftery. Model-based gaussian and non-gaussian clustering. Biometrics, 49(3): 803--821, Sep. 1993.
• 3: C. Cortes, D. Pregibon, and C. Volinsky. Computational Methods for Dynamic Graphs. Journal of Computational and Graphical Statistics, 12(4): 950--970, 2003.
• 4: Valery Guralnik , Jaideep Srivastava, Event detection from time series data, Proceedings of the fifth ACM SIGKDD international conference on Knowledge discovery and data mining, p.33-42, August 15-18, 1999, San Diego, California, United States [doi>10.1145/312129.312190]
• 5: Alexander Ihler , Jon Hutchins , Padhraic Smyth, Adaptive event detection with time-varying poisson processes, Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining, August 20-23, 2006, Philadelphia, PA, USA [doi>10.1145/1150402.1150428]
• 6: Carey E. Priebe , John M. Conroy , David J. Marchette , Youngser Park, Scan Statistics on Enron Graphs, Computational & Mathematical Organization Theory, v.11 n.3, p.229-247, October 2005 [doi>10.1007/s10588-005-5378-z]
• 7: Peter J. Rousseeuw , Katrien Van Driessen, A fast algorithm for the minimum covariance determinant estimator, Technometrics, v.41 n.3, p.212-223, Aug. 1999 [doi>10.2307/1270566]

INDEX TERMS

Primary Classification:
  J. Computer Applications
  J.4 SOCIAL AND BEHAVIORAL SCIENCES

Keywords:
communication networks, event detection