In order to classify meat samples on the basis of turkey adulteration
level, Linear Discriminant Analysis (LDA) was applied to
separated UVevis, NIR and MIR spectra. LDA (Massart et al., 1997)
is a probabilistic classification technique which searches for
directions (canonical variables) with the maximum separation
among categories; the first canonical variable is the direction of
maximum ratio between inter-class and intra-class variances.