Applicability of social network analyses in an intelligence context / Etienne Martineau, DRDC Valcartier. : D68-6/418-2012E-PDF

"This document presents a classification of theoretical social network analyses to determine if they are suitable for the military intelligence context. This work was done under the project 15do, "Social network analysis in a counter insurgency context", to orient the development of a proof of context prototype. Thirty-four (34) theoretical social network analyses were selected and classified based on the implementation risk they pose. The evaluation criteria used are computational complexity, topological and flow requirements and missing data robustness. The risk tolerance of the military intelligence process was also evaluated using the same criteria. Using both results, it was possible to identify 8 suitable algorithms, 11 unusable and 15 risky ones. These results suggest that classical social network analysis methods are not sufficient to answer the military requirements stated in this project. Alternative ways to extract information and perform network analysis should be investigated. In particular, there is a need for algorithms that are robust to missing data and can extract sub-networks of interest efficiently. The impact of missing data is still not well understood and should be the focus of further research as it is largely encountered in the context of covert networks"--Abstract, page i.

Lien permanent pour cette publication :
publications.gc.ca/pub?id=9.876621&sl=1

• Format MARC XML
• Format MARC HTML

Demander des formats alternatifs