Grand Tours for the Selection of Distance 
in High-Dimensional Classification Problems

Jeff Solka, Carey Priebe, Ed Wegman and David Marchette

This talk will focus on our recent work in the selection of distance
measures for high-dimensional classification problems. The selection
of said distance measures are also germane to data clustering. I will
discuss application of grand tour type procedures to the determination
of distance measures optimizing nearest neighbor classification procedures.
The application of these techniques will be illustrated using artificial 
nose hyperspectral data and other hyperspectral imagery.