Statistical methodologies for the collection and analysis of data characterizing large sets of stimuli: Application to the sensory characterization of products and to the characterization of culinary techniques
Méthodologies statistiques pour le recueil et l’analyse de données de caractérisation de grands ensembles de stimuli : Application à la caractérisation sensorielle de produits et à la caractérisation de gestes culinaires
Résumé
In the agri-food sector, the characterization of culinary techniques - through their observation and analysis - is increasingly seen as a lever for innovation. By analyzing how their products are used in the kitchen by the users (consumers or professionals), the manufacturers can detect improvement tracks or ideas for new products. In this research, we propose new methods dedicated to the characterization of culinary techniques. Their innovative nature is based on the fact that they are based on a quantitative approach, and not on a qualitative approach as is usually the case. They partially draw their inspiration in two methods of characterization used in sensory analysis: the free sorting and the binary Q-methodology.
We see how methodological developments brought to the latter make it possible to characterize large sets of stimuli. These developments are both related to the procedure of data collection and to the procedure of statistical analysis of the data. In particular, we propose several statistical procedures to address various issues: the statistical analysis of a set of partitions containing missing data, the unsupervised agreement-based clustering of a set of profiles of binary evaluations, etc. Then, we see that these two 'improved' quantitative methods of characterization can be successfully applicable to culinary techniques.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...