Premium
Methodology to study visual preference for horticultural products using quantitative imaging techniques and latent class segmentation
Author(s) -
Coppenolle Hans,
Paulus I.,
Schrevens E.
Publication year - 2002
Publication title -
international journal of food science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.831
H-Index - 96
eISSN - 1365-2621
pISSN - 0950-5423
DOI - 10.1046/j.1365-2621.2002.00582.x
Subject(s) - pairwise comparison , latent class model , artificial intelligence , segmentation , class (philosophy) , preference , mathematics , conjoint analysis , quality (philosophy) , consistency (knowledge bases) , statistics , computer science , product (mathematics) , pattern recognition (psychology) , philosophy , geometry , epistemology
A methodological concept is proposed to study visual preference for horticultural products using quantitative imaging techniques and latent class segmentation. Chicory, a typical Belgian vegetable, serves as a model product. A 2 3 full factorial pairwise comparison experiment was designed to study visual preference for length, width and ovality of a traditional fresh, high‐quality chicory product using a computer based image system. Twenty respondents chose which of two samples was preferred in all twenty‐eight pairwise combinations of eight constructed image sequences. The consistency of the respondents and agreement between respondents was evaluated. Traditional techniques for analysis of pairwise comparison data were compared with a latent class binomial logit model. This approach relaxes the traditional assumption that the same model holds for all respondents by simultaneously relating choices to quality descriptors for different consumer segments.