CLASSIFICATION OF COMMERCIAL SKIM MILK POWDERS ACCORDING TO HEAT-TREATMENT USING FACTORIAL DISCRIMINANT-ANALYSIS OF NEAR-INFRARED REFLECTANCE SPECTRA

The application of factorial discriminant analysis (FDA) to principal component scores obtained from NIR spectra of a representative set of skim milk powders was attempted to see whether a simple and rapid classification procedure could be developed. Discriminant spectral patterns were produced in an attempt at identifying those molecular groups which were responsible for any discrimination achieved. It is shown that reflectance measurements in the near-infrared spectral region are highly intercorrelated; PC analysis removes this intercorrelation by the creation of new variables (principal components). Thus the application of factorial discriminant analysis on principal components should be more efficient than its use with reflectance data.