A multi-product approach for detecting subjects’ and objects’ covariates in consumer preferences

It is important in the analysis of consumer preference exploits both information deriving from subjects’ and products’ characteristics and this is particularly relevant in Sensometrics where decision to buy depends both on the personal habits and the organoleptic properties of food and beverages. Multivariate analysis copes with these needs by several methods based on factorial approaches and some applications have been performed in order to cluster consumers with respect to products. In this paper a different framework based on a parametric version of the process generating the hedonic scores is adopted. More precisely, a probability distribution for the ordinal responses is proposed as a mixture of feeling and uncertainty components and both of them are related to subjects’ and products’ characteristics. The approach is made effective by inferential methods based on maximum likelihood and asymptotic inference. A real case study is discussed and some advantages are considered.