Characterization and classification of Ohio wines using multivariate data analysis.

摘要

Studies were conducted on characterization and classification on the basis of mutivariate data analysis. 41 wine sampes were used: Chardonnay and Johannisberg Riesling wines from Ohio and California. Data for K, Na, Ca, Cu, Fe, Zn, Mn and Mg contents and 18 aroma compounds were used in the multivariate data analyses. Principal component analysis and K-nearest neighbour analyses were applied. Results showed that wines could be classified by both cv. and geographical origin by this technique; not all the constituents measured were required for discrimination of the wines. Variations attributable to winemaking grape cv. were greater than those attributable to region of production.