Extracting information from sensory analysis with the multi-way PARAFAC tool – Examples and possibilities

Abstract

This study applied an exploratory method named Parallel Factor Analysis (PARAFAC) to extract information from the sensory analysis. Scarce works use PARAFAC on sensory data in scientific literature, and three different ways for data organization and interpretation are shown here. A data set of cachaça sensory analysis showed the possibility of investigating the agreement among trained panelists, similarities between flavored cachaças and those aged in wooden barrels, and how sensory attributes are responsible for the results. Four different coffee beverage types were distinct by the roasted degree related to the beverage quality. The agreement among sensory results from the same beverage type was pointed as precision measurement at repeatability level. For cheese bread formulations with sweet potato flour, the considered attributes highlighted the formulation most acceptable for the panelists. They can help evaluate male and female preferences or the consumer's age.