Objective evaluation of various trade varieties of coffee by coupling of analytical data and multivariate analyses

The aroma profiles of 39 coffee samples (32 arabica and 7 robusta coffees) were evaluated objectively by the coupling of gas chromatographic analysis and two kinds of multivariate analysis (principal component analysis and cluster analysis) and compared with the classification on the basis of a cup test of brewed coffee by cup testers. Robusta coffees were separated from arabica coffees by principal component analysis (PCA) and cluster analysis. Using PCA of only 32 arabica coffee samples, their aroma profiles could be characterized on the first and second principal components. The components responsible for grassy aroma and earthy odor, respectively, were clarified from the factor loadings obtained by PCA. Furthermore, 32 arabica coffees were divided into seven clusters by cluster analysis of the first and second principal components obtained by PCA. Consequently, 39 coffee samples were classified into eight groups and the result was consistent with the classification on the basis of sensory evaluation except for two samples.