Discrimination of beer flavours by analysis of volatiles using the mass spectrometer as an electronic nose

Abstract

Entire mass spectra of beer headspace components were used as fingerprints for beer brand classification and differentiation of beer samples stored under various conditions. Chemometric analysis of the mass spectra allowed for the discrimination of beer brands and for the detection of beer aging and photodegradation. The numeric methods used include unsupervised PCA modelling and discrimination using kNN, LDA, and D-PLS methods. © 2007 The Institute of Brewing & Distilling.