With recent terrorist actions such as the anthrax delivery through the US Postal Service, there is an urgent need for an early warning system for disease outbreaks. Charles River Analytics has created the Automated Link Analysis for Data-mining of Distributed Information (ALADDIN) Early-Warning System for Epidemics or Bio-terror Attacks. ALADDIN is a system employing a Bayesian Belief Network approach for such early detection.
Belief networks are powerful modeling tools for condensing what is known about causes and effects into a compact network of probabilities. ALADDIN uses Charles River Analytics’s Bayesian Belief Network engine, BNet.
The technical approach of ALADDIN is to fuse data from multiple sources and generate alerts. It discovers non-obvious connections between entities to further establish evidence of a potential attack. ALADDIN employs a model to deal with the heterogeneous nature of data sources. It has validated its approach by simulating bioterrorist attacks in model cities.
For automated link discovery, Charles River Analytics is using the Strategist link discovery packages from its partner, Electronic Warfare Associates. Charles River Analytics is also using Carnegie Mellon University’s BioWar simulation environment to support what-if analysis and model-generated hypothesis.