PRINCIPAL COMPONENTS AND CLUSTER-ANALYSIS FOR DESCRIPTIVE SENSORY ASSESSMENT OF INSTANT COFFEE

The relationships among 13 aroma, flavor, mouthfeel and appearance var iables for 18 soluble coffees were analyzed using flavor profiling. Th ree-way ANOVA showed significant main effects for coffees and judges i n all attributes. The data were submitted to principal component analy ses (PCA) and cluster analysis (CA). Two sequential PCA were performed . The first PCA showed that flavor, bitterness and duration were the m ost important descriptors positively correlated with the first PC, whi le the variation in appearance properties dominated the second PC, neg atively correlated with these attributes. Five attributes were elimina ted and a subset of 8 variables was submitted to a second PCA. The mea ning of the first two PC remained unchanged and, as expected, the tota l variation explained by the first four PC increased. Frequency of pos itive and negative judgments in both PC allowed to separate coffees in to four categories. Confirming the choice of the variables, the CA rev ealed similar distribution Of coffees into four clusters. Aroma, flavo r and mouthfeel attributes seemed to play a more important role in the determination of clusters than the appearance variables.