TRANSMet™ : Human Network Behavior Analysis and Prediction

Human network analysis using traditional network metrics tends to be impoverished when it comes to understanding or predicting the behavior of the human network. To overcome this limitation, requires the addition of behavioral semantics to the human nodes in the network. Specifically, there is a need for defining and computing novel network metrics which can be translated into behavioral attributes that can be associated with human nodes as behavioral metadata and visualized in actionable form. This SBIR effort is concerned with developing an analysis engine that can compute network metrics from raw network data, identify applicable behavior attributes from the network metrics, and update node descriptors with new behavioral metadata on an ongoing basis. With this capability in place, it becomes ultimately possible to predict the response of the human network to different stimuli. Phase I of this effort is intended to establish the feasibility and tractability of the overall approach with arbitrary data sets that include ground truth.