Nonnegative Matrix and Tensor Factorization for Discussion Tracking - Text Mining: Classification, Clustering, and Applications - page 92

Chapter 5

Nonnegative Matrix and Tensor
Factorization for Discussion Tracking

Brett W. Bader, Michael W. Berry, and Amy N. Langville

5.1 Introduction ............................................................. 95

5.2 Notation ................................................................. 97

5.3 Tensor Decompositions and Algorithms ................................. 98

5.4 Enron Subset ............................................................ 102

5.5 Observations and Results ................................................ 105

5.6 Visualizing Results of the NMF Clustering ............................. 111

5.7 Future Work ............................................................. 116

5.1 Introduction
After the filing for Chapter 11 bankruptcy by Enron in December of 2001,
an unprecedented amount of information (over 1.5 million electronic mail mes-
sages, phone tapes, internal documents) was released into the public domain.
Such information served the needs of the Federal Energy Regulatory Commis-
sion (FERC) in its investigation against Enron. The emails originally posted
on the FERC web site (18) had various integrity problems which required some
cleaning as well as the removal of sensitive (private) and irrelevant informa-
tion. Dr. William Cohen and his research group at Carnegie Mellon University
have addressed many of these problems in their release of the Enron Email
Sets. The version of the Enron Email Sets 1 dated March 2, 2004 contains
517 , 431 email messages of 150 Enron email accounts covering a period from
December 1979 through February 2004 with the majority of messages span-
ning the three years: 1999, 2000, and 2001.
The emails in this corpus reflect the day-to-day activities of what was the
seventh largest company in the United States at that time. There were,
however, certain topics of discussion uniquely linked to Enron activities (5).
Enron's development of the Dabhol Power Company (DPC) in the Indian

1 http://www-2.cs.cmu.edu/ ~ enron