Prediction of the sensory acceptance of fruits by physical and physical-chemical parameters using multivariate models

Data about overall liking and physical and physical-chemical analysis for oranges, pineapples, and grapes were analyzed by Principal Component Analysis (PCA). Results showed that solid soluble variables, soluble solids content/total titratable acidity ratio, and pH contributed positively and titratable acidity contributed negatively to the overall liking grade, indicating preference for sweeter and less acidic fruit samples. Consumer acceptances were calibrated against physical and physical-chemical measurements of those fruits using Multiple Linear Regression. The models obtained were then validated and tested using the widely used methods of y-randomization and external validation. In all cases, multivariate models presented R2 values >0.7, which were higher than for the univariate models. Therefore, the models built and validated for oranges, pineapples, and grapes can be used to predict the consumer acceptance by easy and quick physical and physical-chemical measurements, ensuring that fruit commercialization takes sensory acceptance into consideration.