Project quad chart. (Powerpoint)

Algorithms for anomaly, intrusion and masquerade detection have been thus far developed and evaluated in an ad hoc manner. The performance of these algorithms cannot be predicted in datasets other than the ones on which they were tested; sometimes their performance cannot even be replicated. The field of information assurance needs better quality and higher predictability. The objective of the Profiler-2000 project is to improve detection performance by:

- Developing a basic science of profiling.

- Developing a diverse suite of detectors.

- Providing custom, calibrated testbeds.

- Providing statistically and methodologically rigorous assessment procedures.