Grade classification of canned pink salmon with static headspace volatile patterns

Pacific pink salmon (Oncorhynchus gorbuscha) were subjected to refrigerated decomposition prior to canning. Data on 44 volatile compounds obtained by static headspace gas chromatography were analyzed using multivariate statistical methods. Principal component analysis and common factor analysis showed two dimensions were particularly important to delineate variation between quality grades A, B, and reject. Three types of discriminant analysis (DA) were carried out on PCA scores as well as three selected volatiles (ethanol, 3-methyl-1-butanol, and 2-methyl-l-propanal). Although all DA procduced misclassification rates below 7%, the nonparametric Epanechnikov kemal DA was more reliable followed by quadratic and linear DA. Use of 3-methyl-1-butanol and 2-methyl-1-propanal in addition to ethanol improved the detection of initial decomposition stages.