APPLICATION OF MULTIVARIATE STATISTICAL ANALYSIS FOR THE QUALITY CONTROL OF FOOD PRODUCTS
Special Symposium - Innovations in Food Technology (LMC Congress)
Flexible Production, PAT & Modelling (Food-3b)
Keywords: Multivariate analysis, quality control, classification, cheese
A proposal for the quality control of food products by using multivariate statistical analysis is described. Multivariate statistical methods comprising principal component analysis (PCA), cluster analysis (CA) and stepwise discriminant analysis (SDA) were applied to estimate the usefulness of various chemical, physical and sensory determinations for the differentiation and classification of the different types of Majorcan cheese: half ripened, ripened and old ripened. Chemical composition (dry matter and non-protein nitrogen) and physical parameters (water activity, yellowness colour index and the texture parameters hardness, cohesiveness and elasticity) have been measured. Sensory evaluation of odour and aroma intensity, elasticity, firmness, friability, olfactory and gustative sensations (aftertaste) and persistence has been carried out by 11 trained judges. PCA has reduced chemical, physical and sensory variables to three independent components accounting for the 85.7% of the explained variation, representing the first principal component the ripening stage. CA has confirmed the correlation between the studied variables obtained by PCA. SDA has determined which variables best classify cheese samples according to their ripening stage, being able to differentiate between the three stages of ripening suggested by the manufacturers. Using classification functions obtained by SDA, 93.5% of cheese samples were correctly classified according to their ripening stage. Therefore, multivariate statistical methods comprising principal component analysis (PCA), cluster analysis (CA) and stepwise discriminant analysis (SDA) have shown to be effective in order to select the optimal parameters for the quality control of Majorcan cheeses; moreover cheeses could also be correctly classified according to their ripening stage.
Presented Wednesday 19, 17:30 to 17:35, in session Flexible Production, PAT & Modelling (Food-3b).
