[FoRK] Machine learning at work... Enron email

Jim Whitehead < ejw at soe.ucsc.edu > on > Mon Jun 12 09:20:13 PDT 2006


A whitepaper presenting an analysis of Enron email. Describes many  
examples of personal email sent using a company email address,  
including one night stands, sexist jokes, racial stereotypes, and more.

The whitepaper is by a company that sells software to automatically  
detect policy violations in corporate email -- a classic machine  
learning classification problem. Based on the data they present, it  
looks like you could be credible in this field just by taking the  
Support Vector Machine classifier from the WEKA toolkit and applying  
it to a corporate email database (after training, of course).

You can sign up for a free InBoxer account and search the Enron email  

- Jim 

More information about the FoRK mailing list