Spectral stress strain analysis and partial least squares regression to predict sensory texture of yogurt using a compression/penetration instrumental method

The sensory texture characteristics of 2 yogurt types (light, blended, N = 69) were evaluated using a trained descriptive panel and an instrumental compression/penetration test in combination with a novel data analysis method (that is Spectral Stress Strain Analysis). Partial Least Squares Regression was used to study the relationship between each of the 7 sensory texture attributes and spectral force deformation data measured during instrumental testing. The best predictive models were computed for spoon impression (Validation Correlation (R-val) = 0.93, Root Mean Square Error of Prediction (RMSEP) = 1.33, Discrimination Index (DI) = 2.80), visual thickness (R-val) = 0.89, RMSEP = 0.83, DI = 2.40), and slipperiness (R-val = 0.81, RMSEP = 0.60, DI = 1.76). Oral Thickness (R-val = 0.78), cohesiveness (R-val = 0.73), covering (R-val = 0.61) and stickiness (R-val = 0.57) were less accurately predicted.