文摘
We report the evaluation of the floral origin of honey by analysis of its volatile organic compounds(VOCs) profile, joined with the use of combined pattern recognition techniques. Honey samples, fromfive floral origins, were analyzed by headspace solid-phase microextraction-gas chromatography-mass spectrometry, selecting 35 VOCs out of the entire profiles, which were analyzed by hierarchicalcluster analysis (HCA), stepwise discriminant analysis (SDA), and K-nearest-neighbor (KNN). BothHCA and SDA were used as exploratory tools to select a group of VOCs representing similitude anddifferences among studied origins. Thus, six out of 35 VOCs were selected, verifying theirdiscriminating power by KNN, which afforded 93% correct classification. Therefore, we drasticallyreduced the amount of compounds under consideration but kept a good differentiation between floralorigins. Selected compounds were identified as octanal, benzeneacetaldehyde, 1-octanol, 2-methoxyphenol, nonanal, and 2-H-1-benzopyran-2-one. The analysis of VOC profiles, coupled to HCA,SDA, and KNN, provides a feasible alternative to evaluate the botanical source of honey.