Application of Social Network Analysis Techniques to Machine Translated Documents

Report No. ARL-MR-741
Authors: Ann E. M. Bornstein, John H. Brand, Michelle C. McVey, and Sean Murray
Date/Pages: April 2010; 60 pages
Abstract: This report documents and closes out the investigation into use of machine translation (MT) to provide material to generate actionable intelligence. A capability for automatic, real, or near real-time extraction of information on the local and external social networks will help produce actionable intelligence while the forces are at the site and able to act on the intelligence. Freeware tools exist to perform extraction of social network information from textual material. Using a social network analysis (SNA) toolset on textual material may provide the insights into social and, for irregular forces, command and action hierarchies while the area and its inhabitants are under friendly control. On-site exploitation of textual material surfaced during the SNA process may allow disruption of enemy cellular organization and logistic support assets and networks. SNA of this textual material thus constitutes an application of data mining in support of tactical operations. The SNA and information exploitation enablers teams conducted an investigation to determine the degree to which MT preserves concept maps, including social network maps. A freeware SNA toolset was applied to a series of texts translated by various MT engines. The SNA tools were exercised on a set of MT documents provided by the U.S. Army Research Laboratory Multilingual Computing Research Branch and on documents acquired from the Internet. This report documents results of the investigation.
Distribution: Approved for public release
  Download Report ( 1.001 MBytes )
If you are visually impaired or need a physical copy of this report, please visit and contact DTIC.
 

Last Update / Reviewed: April 1, 2010