Identification of Incomplete Networks

Sponsor: Office of Naval Research

PI: James E. Driskell, Ph.D.

International terrorist networks are characterized by dynamic, cell-based organizational networks.  Understanding and predicting their capabilities and resources is a critical defense requirement, and useful tools based on social network analysis have been developed to identify the structure and dynamics of these networks in order to control and contain their activities.  However, traditional social network analysis approaches typically assume near-complete and full information regarding a network, whereas available information about covert, terrorist networks is often based upon missing and incomplete data.  Existing analysis tools and techniques do not account for missing or incomplete data, and experts have warned that ignoring the problem of incomplete data can lead to erroneous and possibly disastrous consequences.

This research project examines a novel approach for estimating the size of elusive or covert populations based on the capture-recapture methodology.  Capture-recapture techniques have been used in various fields such as epidemiology to estimate the size of populations that are difficult to count or are highly mobile.  This research entails an adaptation of the capture-recapture method to be employed with social network data. This approach provides a means to gauge the fidelity of the data used to construct terrorist social networks and to provide an estimate of the actual total size of the terrorist network. This product is useful for analysts to determine whether, and how many, as-yet unidentified members of a terrorist network exist.  This product can be adapted to any current network analysis tool to improve analysis of critical terrorist networks.