Multiblock modeling for complex preference study. Application to European preferences for smoked salmon

Type Article
Date 2014-03
Language English
Author(s) Bougeard Stephanie1, Cardinal Mireille2
Affiliation(s) 1 : Anses French Agcy Food Environm & Occupat Hlth Sa, Dept Epidemiol, F-22440 Ploufragan, France.
2 : IFREMER, Lab Sci & Technol Biomasse Marine, F-44311 Nantes 03, France.
Source Food Quality And Preference (0950-3293) (Elsevier Sci Ltd), 2014-03 , Vol. 32 , N. Part A , P. 56-64
DOI 10.1016/j.foodqual.2013.06.002
WOS© Times Cited 4
Keyword(s) Multiblock modeling, Multiblock redundancy analysis, External preference mapping, Smoked salmon
Abstract The aim of the paper is to propose an alternative method to external preference mapping for the case of complex data where explanatory variables are organized in meaningful blocks. We propose an innovative method in the multiblock modeling framework, called multiblock Redundancy Analysis. The interest and relevance of this method is illustrated on the basis of a European consumer preference study for cold-smoked salmon. The study aims at explaining six homogeneous clusters of preference with explanatory parameters organized in five thematic blocks related to physico-chemical measurements, microbiological characterization, appearance attributes, odor/flavor characterization and texture descriptors. Overall indexes and graphical displays associated with different interpretation levels are proposed to sort the key drivers of preference by order of priority at the variables and at the block level. On the basis of these data, multiblock Redundancy Analysis is also compared to standard preference mapping in terms of model quality; the best model is here associated with the multiblock method.
Full Text
File Pages Size Access
9 1 MB Access on demand
Author's final draft 17 941 KB Open access
Top of the page