An information theoretic comparison of projection pursuit and principal component features for classification of Landsat TM imagery of central Colorado

Projection pursuit (PP) and principal component analysis(PCA) projections d erived from Landsat Thematic Mapper (TM) imagery of central Colorado were c ompared. While PCA is a simple subset of the general class of PP algorithms , it cannot distinguish Gaussian from non-Gaussian distributions, since it maximizes projected variance. PP algorithms, which maximize higher-order st atistics, can be used to find skew or multi-modal projections in order to r eveal underlying class structure. These data projections have greater fidel ity to underlying land-cover distributions. On sequestered test data, PP pr ojections improved separation of individual categories from a few percent t o as much as 24%. PP performance exceeded that of PCA for all but one of th e 14 land-cover categories.