When multiple measurements are made on a set of samples by wet chemical, spectral, electrophoretic or chromatographic methods, or by combinations thereof, it is possible to use multivariate pattern recognition to establish sample classifications. These classifications can be by the growing location, crop year, or cultivar of an agricultural commodity, or by the brand or production plant of a processed food product. Methods such as principal components analysis (PCA), linear discriminant analysis (LDA), K-nearest neighbor analysis (KNN) or SIMCA may then be used to classify samples or to detect mis-matches in any direction in a multivariate sense. Adulterated or misrepresented samples can often be identified in this way. Examples using hop essential oils, beer volatile compounds and combinations of various brewing analytical methods will be shown. Advantages and limitations of various methods will be described.

Using chemometrics to classify samples and detect misrepresentation

239th National Meeting (2010)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.