Events and Meetings of Italian Statistical Society, Advances in Latent Variables - Methods, Models and Applications

Font Size: 
Uncovering the latent structure in sensory data
El Mostafa Qannari

Last modified: 2013-06-16


Since sensory analysis fully stood out as a science some three or four decades ago, there have been many developments both from conceptual and technical standpoints. These developments would have not been possible without sensometrics methods to obtain relevant information from (very often) noisy data. We review some typical problems is sensory evaluation namely, sensory profiling, paired comparison data and free sorting task, and we highlight statistical approaches to analyze the data. We focus on statistical approaches based on the determination of latent variables (or components) that reflect the main differences among the products being assessed.

Full Text: PDF