Physics-based feature mining for large data exploration

One effective way of exploring large scientific data sets is a process called feature mining. The two approaches described here locate specific features through algorithms that are geared to those features underlying physics. Our intent with both approaches is to exploit the physics of the problem at hand to develop highly discriminating, application-dependent feature detection algorithms and then use available data mining algorithms to classify, cluster, and categorize the identified features. We have also developed a technique for denoising feature maps that exploits spatial-scale coherence and uses what we call feature preserving wavelets. The examples presented demonstrate our feature mining approach as applied to steady computational fluid dynamics simulations on curvilinear grids.