The Air Force has a need for intelligent tools that can be used to convert information from biodynamics databases into knowledge and decisions. Current methods of feature extraction and hypothesis testing require significant amounts of human interpretation. The innovative techniques offered in this proposal utilize feature-independent and automated methods to facilitate scientific advancement. The resulting intelligent hypothesis testing tool can increase the rate and exploration of data mining, analysis, and feature extraction for data fusion. The proposed Phase I research and development seeks to construct algorithms that optimize hypotheses through feature extraction. Evolutionary computing is used to find optimal representations relating database features to predictions of outcome. The algorithms will be designed for use with Air Force biodynamics databases. The Phase I research and development sets the stage for continued Phase II research and development and transition for field use. The technology's applications go beyond Air Force database analysis to all branches of the military, and also commercial and academic database analysis particularly in bioinformatics. The prospect for commercialization for the resulting technology in the bioinformatics sector is high in light of the fact that database mining is a common facet of gene expression analysis and drug discovery.
