Advertisement

Computational & Mathematical Organization Theory

, Volume 11, Issue 3, pp 249–264 | Cite as

Email Surveillance Using Non-negative Matrix Factorization

  • Michael W. Berry
  • Murray Browne
Article

Abstract

In this study, we apply a non-negative matrix factorization approach for the extraction and detection of concepts or topics from electronic mail messages. For the publicly released Enron electronic mail collection, we encode sparse term-by-message matrices and use a low rank non-negative matrix factorization algorithm to preserve natural data non-negativity and avoid subtractive basis vector and encoding interactions present in techniques such as principal component analysis. Results in topic detection and message clustering are discussed in the context of published Enron business practices and activities, and benchmarks addressing the computational complexity of our approach are provided. The resulting basis vectors and matrix projections of this approach can be used to identify and monitor underlying semantic features (topics) and message clusters in a general or high-level way without the need to read individual electronic mail messages.

Keywords

electronic mail Enron collection non-negative matrix factorization surveillance topic detection constrained least squares 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Berry, M. and M. Browne (2005), Understanding Search Engines: Mathematical Modeling and Text Retrieval (2nd ed.). Philadelphia, PA: SIAM.Google Scholar
  2. Berry, M., Z. Drmač, and E. Jessup (1999), “Matrices, Vector Spaces, and Information Retrieval,” SIAM Review, 41(2), 335–362.Google Scholar
  3. Donoho, D. and V. Stodden (2003), “When does Non-Negative Matrix Factorization Give a Correct Decomposition into Parts?” Technical report,Department of Statistics, Stanford University. Preprint.Google Scholar
  4. Giles, J., L. Wo, and M. Berry (2003), “GTP (General Text Parser) Softwarefor Text Mining,” in H. Bozdogan (Ed.), Software for Text Mining, in Statistical Data Mining and Knowledge Discovery, Boca Raton,FL: CRC Press, pp. 455–471.Google Scholar
  5. Grieve, T. (2003, October 14). The Decline and Fall of the Enron Empire.Slate. http://www.salon.com/news/feature/2003/10/14/enron/index_np.html.
  6. Guillamet, D. and J. Vitria (2002), “Determining a Suitable Metricwhen Using Non-Negative Matrix Factorization,” in Sixteenth International Conference on Pattern Recognition (ICPR'02), Vol. 2, Quebec City, QC, Canada.Google Scholar
  7. Hoyer, P. (2002), “Non-Negative Sparse Coding,” in Proceedings of the IEEEWorkshop on Neural Networks for Signal Processing, Martigny, Switzerland.Google Scholar
  8. Hyvärinen, A. and P. Hoyer (2000), “Emergence of Phase and Shift Invariant Features by Decomposition of Natural Images into Independent Feature Subspaces,” Neural Computation, 12(7), 1705–1720.Google Scholar
  9. Jolliffe, I. (2002), Principle Component Analysis (2nd ed.).New York: Springer-Verlag.Google Scholar
  10. Keila, P. and D. Skillicorn (2005, April 23), “Structure in the Enron Email Dataset,” in Proceedings of the Link Analysis, Counterterrorism, and Security Workshop, Fifth SIAM International Conference on Data Mining, Newport Beach, CA, pp. 55–64.Google Scholar
  11. Lee, D. and H. Seung (1999), “Learning the Parts of Objects by Non-Negative Matrix Factorization,” Nature, 401, 788–791.Google Scholar
  12. Lee, D. and H. Seung (2001), “Algorithms for Non-Negative Matrix Factorization,” Advances in Neural Information Processing Systems, 13, 556–562.Google Scholar
  13. Liu, W. and J. Yi (2003), “Existing and New Algorithms for Non-Negative Matrix Factorization,” Technical report, Departmentof Computer Sciences, University of Texas at Austin. Preprint.Google Scholar
  14. McCallum, A., A. Corrada-Emmanuel, and X. Wang (2005, April 23). “The Author-Recipient-Topic Model for Topic and Role Discovery in Social Networks with Application to Enron and Academic Email,” in Proceedings ofthe Link Analysis, Counterterrorism, and Security Workshop, Fifth SIAMInternational Conference on Data Mining, Newport Beach, CA, pp. 33–44.Google Scholar
  15. McLean, B. and P. Elkind (2003), The Smartest Guys in the Room: The Amazing Rise and Scandalous Fall of Enron. Portfolio.Google Scholar
  16. Mu, Z., R. Plemmons, and P. Santago (2003), “Iterative Ultrasonic Signaland Image Deconvolution for Estimating the Complex Medium Response,”in IEEE Transactions on Ultrasonics and Frequency Control, IEEE, Submitted for publication.Google Scholar
  17. Prasad, S., T. Torgersen, V. Pauca, R. Plemmons, and J. van der Gracht(2003), “Restoring Images with Space Variant Blur via Pupil Phase Engineering,” Optics in Info. Systems, Special Issue on Comp. Imaging, SPIEInt. Tech. Group Newsletter, 14(2), 4–5.Google Scholar
  18. Shahnaz, F., M. Berry, V. Pauca, and R. Plemmons (2006), “Document Clustering Using Nonnegative Matrix Factorization,” Information Processing and Management, 42(2), 373–386.CrossRefGoogle Scholar
  19. Xu, W., X. Liu, and Y. Gong (2003), “Document-Clustering based on Non-negative Matrix Factorization,” in Proceedings of SIGIR'03, July 28–August 1, Toronto, CA, pp. 267–273.Google Scholar

Copyright information

© Springer Science + Business Media, Inc. 2006

Authors and Affiliations

  1. 1.Department of Computer ScienceUniversity of TennesseeKnoxville

Personalised recommendations