Estimation of adversarial social networks by fusion of information from a wide range of sources

A data structure is described that serves to define a target structure for estimating social networks. It represents who knows whom, the strength and polarity of associations, and levels of confidence that linked individuals are actually personally acquainted. This network, which represents social structure, is embedded in a more inclusive network structure that also represents vehicles, places and organizations. This broader structure can be used to accumulate information to enable automated inferencing to assist human processing in estimating the social network of interest. How portions of this inferencing can be done is briefly described, and a scenario is given that illustrates the use of this kind of fused information to make a decision. We also review a range of hard and soft information sources with emerging value for estimating adversarial social networks, and describe how these sources can be used for the purpose.