Automated Link Analysis for Data-mining of Distributed INformation (ALADDIN)

"We propose to develop a system for Automated Link Analysis for Data-mining of Distributed INformation (ALADDIN) to detect bioterrorist attacks. ALADDIN will use ontology-based data modeling for integrating heterogeneous and distributed data sources ofbioterrorism information. In addition to traditional sources such as hospital and ambulatory-care records, ALADDIN will incorporate data from other potentially relevant sources, such as intelligence agencies, Customs databases, and INS records. ALADDINwill include a manual link discovery tool to support the investigation of inter-person or inter-organizational links between otherwise autonomous data sources (e.g., linking the INS file of a suspected terrorist with a Custom's Service record of thatperson importing a shipment of hazardous chemicals). ALADDIN will also incorporate Bayesian belief network (BN) technology for automatically identifying data sources for potential leads. BN models will be used to represent interrelationships among variousdomain relevant concepts such as terrorist activities, diseases, and symptoms, thus allowing for the integration of both clinical and non-clinical information in the detection of potential bioterrorism attacks. We intend to rapidly develop a prototypebased on: 1) a domain ontology encoded with Prot¿g¿ 2000 from SMI; 2) the link analysis tool Strategist from our strategic business partner Electronic Warfare Associates (EWA); and 3) our in-house B