Using contrasts as data pretreatment method in pattern recognition of multivariate data

A contrast method originally proposed by Spiegelman [C.H. Spiegelman, Calib ration: a look at the mix of theory, methods and experimental data, present ed at Compana '95, Wuerzburg, Germany.] is modified to pretreat multivariat e data for classification. Three NIR data sets and one pollution data set a re used as examples. Our results show that the contrast method greatly impr oves the ratios of between- to within-class variance. It is more powerful t han offset correction, SNV, first- and second-derivative methods in the cas es studied. This conclusion does not depend on the type of classifier used. Regularised discriminant analysis (RDA) and partial least squares (PLS2) w ith univariate feature selection based on Fisher's ratio were applied here. There is a risk that chance correlations occur after the contrast pretreat ment. The chance correlation decreases after first eliminating un-informati ve variables using the modified Uninformative Variable Elimination (UVE)-PL S method. (C) 1999 Elsevier Science B.V. All rights reserved.