文摘
The volatile congener analysis of 52 commercialized whiskeys (24 samples of single malt Scotchwhiskey, 18 samples of bourbon whiskey, and 10 samples of Irish whiskey) was carried out by gaschromatography/mass spectrometry after liquid-liquid extraction with dichloromethane. Patternrecognition procedures were applied for discrimination of different whiskey categories. Multivariatedata analysis includes linear discriminant analysis (LDA), k nearest neighbors (KNN), soft independentmodeling of class analogy (SIMCA), procrustes discriminant analysis (PDA), and artificial neuralnetworks techniques involving multilayer perceptrons (MLP) and probabilistic neural networks (PNN).Classification rules were validated by considering the number of false positives (FPs) and falsenegatives (FNs) of each class associated to the prediction set. Artificial neural networks led to thebest results because of their intrinsic nonlinear features. Both techniques, MLP and PNN, gave zeroFPs and zero FNs for all of the categories. KNN is a nonparametric method that also provides zeroFPs and FNs for every class but only when selecting K = 3 neighbors. PDA produced good resultsalso (zero FPs and FNs always) but only by selecting nine principal components for class modeling.LDA shows a lesser classification performance, because of the building of linear frontiers betweenclasses that does not apply in many real situations. LDA led to one FP for bourbons and one FN forscotches. The worse results were obtained with SIMCA, which gave a higher number of FPs (five forboth scotches and bourbons) and FNs (six for scotchs and two for bourbons). The possible cause ofthese findings is the strong influence of class inhomogeneities on the SIMCA performance. It isremarkable that in any case, all of the methodologies lead to zero FPs and FNs for the Irish whiskeys.Keywords: Discrimination; whiskey; pattern recognition