The utilization of Bayesian statistics in the realm of food sciences

Abstract

In Bayesian statistics, even without objective data (prior probabilities), estimations can be formulated by subjectively defining prior probabilities. Probabilities (posterior probabilities) change as fresh data are acquired, and reasonable inferences can be drawn from a limited dataset. Moreover, inferences become more precise with larger amount of data. In the field of food quality and safety management, numerous variables remain unobserved or unmeasured due to economic or technical limitations. Bayesian networks are highly suitable for addressing these variables and can be employed to construct models for diverse aspects of quality control and the identification of potential hazards. Furthermore, they possess the capability to handle uncertainty and prove valuable in product design by offering a means to comprehensively approach intricate issues. Bayesian networks are garnering growing interest, especially in food-related consumer sensory and marketing investigations, and further applications are anticipated in this domain.