How to learn consumer preferences from the analysis of sensory data by means of support vector machines (SVM)


In this paper, we discuss how to model preferences from a collectionof ratings provided by a panel of consumers of somekind of food product. We emphasize the role of tasting sessions,since the ratings tend to be relative to each sessionand hence regression methods are unable to capture consumerpreferences. The method proposed is based on the use of SupportVector Machines (SVM) and provides both linear andnonlinear models. To illustrate the performance of theapproach, we report the experimental results obtained witha couple of real world data sets.