Application of agglomerative hierarchical clustering to identify consumer tomato preferences: influence of physicochemical and sensory characteristics on consumer response

A multiple regression model was developed to predict the acceptability of the four tomato varieties studied (Aranka, Cherry, Beef and Pitenza). Agglomerative hierarchical clustering showed the presence of four consumer clusters. One cluster preferred small tomatoes (Aranka and Cherry) and another cluster the larger tomatoes (Beef and Pitenza). In the sensorial analysis Aranka was the preferred variety, scoring more highly in taste, odour, acidity, sweetness and hardness. In the physicochemical analysis Aranka also obtained the highest values for titratable acidity (TA) and sugars (SSC), confirming that these parameters are important in tomato flavour. Lower values for both sets of parameters were reflected by lower consumer acceptability, with Beef and Pitenza receiving the lowest score for these flavours attributes (except odour). A significant correlation between the sensorial and physicochemical parameters was also observed: odour was positively correlated with calibre, while taste, acidity, hardness and acceptability were negatively correlated with calibre, pH and SSC/TA and positively correlated with SSC and TA. Copyright © 2005 Society of Chemical Industry