A Low Cost, Hybrid Approach to Data Mining, Phase I

The proposed effort will combine a low cost physical modeling approach with inductive, data-centered modeling in an aerosopace relevant context to demonstrate effective, low cost data mining. In particular Phase I will evaluate various hybrid architecture concepts on the basis of false positive and fasle negative rates. The approach will use domain decompostiition to partition the physical platform under consideration into regimes appropriate for either model based or inductive based apoproaches. More »