Comparison of classical kinetic and Bayesian approaches for predicting food shelf life by accelerated testing: the case of biobased coffee capsules

Predicting the shelf life (SL) of food products is still challenging from a computational point of view. In the present work, the classical kinetic (K) approach was compared with the Bayesian (B) methodology, whose application in SL studies is largely unexplored. Data relevant to pH changes of coffee brews obtained from coffee packaged in bio-based capsules and stored at different relative humidity (54, 65, 75 %) and temperature (20, 30, 45 °C) were considered as a case study, assuming pH 5.1 as the acceptability limit of the coffee brew. Data were elaborated according to K or B methodologies and SL estimates were compared. Although B better fit coffee pH decay at all environmental conditions, this methodology provided SL estimates comparable to those obtained by K. However, the B methodology produced SL estimates with considerably smaller uncertainty intervals and gave the opportunity to interpret shelf life from a probabilistic point of view.