Monitorization of consumer and naif panels in the sensory evaluation of two types of potato chips by predictive biplots applied to generalized procrustes and three-way Tucker-1 analysis

• Tipo: Artigo em Revista Científica Internacional
• Data: 2005

Autores

Alves, MR (Autor) Outra; Oliveira, MB (Autor) FFUP

Revista

Título: Journal of Chemometrics

Vol. 19 Nº 10

Páginas: 564-574

ISSN: 0886-9383

Editora: Wiley-Blackwell

Indexação

ISI Web of Knowledge - 3 Citações

ISI Web of Science

Scopus - 4 Citações

Classificação Científica

CORDIS: Ciências da Saúde

FOS: Ciências da engenharia e tecnologias > Engenharia electrotécnica, electrónica e informática

Outras Informações

ID Authenticus: P-000-14J

DOI: 10.1002/cem.963

Resumo (PT): Potato chips of industrial and domestic types, produced in batch and continuous plants, with oils of different qualities, were analyzed by a naïf panel with five assessors following an insipient quantitative descriptive analysis (QDA), and by a panel of 15 consumers following a simplified free-choice profiling (FCP) technique. The naïf panel results were analyzed by a three-way Tucker-1 common loadings model with predictive biplots, which proved to be a very good way to evaluate each assessor's perceptions and scoring patterns and to monitor the evolution of the overall panel's performance during training, including the validation of the list of attributes. The consumer panel results were also analyzed by predictive biplots applied to individual dimensions in generalized procrustes analysis (GPA), following data compression through principal components analysis. This technique, which can be extended to a detailed analysis of scoring patterns, enabled the checking of the main consumer's perceptions and beliefs, and simultaneously the evaluation of the interplay of constructs in the build up of consumers' main dimensions. Both panels were able to distinguish the main types of potato chips, as well as the different status of oil quality, producing similar overall solutions to the problem. However, it is observed that when working with consumer panels, due to the great variability involved, at least a reasonably trained sensory panel will help to validate the results obtained <br> <br> Keywords : Generalized procrustes analysis • three-way analysis • predictive biplots • sensory analysis • free-choice profiling • QDA <br> <a target="_blank" href="http://www3.interscience.wiley.com/journal/112602335/abstract"> Texto integral </a> <br> <br>

Abstract (EN): Potato chips of industrial and domestic types, produced in batch and continuous plants, with oils of different qualities, were analyzed by a naif panel with five assessors following an insipient quantitative descriptive analysis (QDA), and by a panel of 15 consumers following a simplified free-choice profiling (FCP) technique. The naif panel results were analyzed by a three-way Tucker-1 common loadings model with predictive biplots, which proved to be a very good way to evaluate each assessor's perceptions and scoring patterns and to monitor the evolution of the overall panel's performance during training, including the validation of the list of attributes. The consumer panel results were also analyzed by predictive biplots applied to individual dimensions in generalized procrustes analysis (GPA), following data compression through principal components analysis. This technique, which can be extended to a detailed analysis of scoring patterns, enabled the checking of the main consumer's perceptions and beliefs, and simultaneously the evaluation of the interplay of constructs in the build up of consumers' main dimensions. Both panels were able to distinguish the main types of potato chips, as well as the different status of oil quality, producing similar overall solutions to the problem. However, it is observed that when working with consumer panels, due to the great variability involved, at least a reasonably trained sensory panel will help to validate the results obtained. Copyright (C) 2006 John Wiley & Sons, Ltd.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 11