“…Both the logistic-curve and logistic-regression models are so-call parametric models in which all the parameters of interest are in the finite-dimensional parameter spaces. With the development of big data technologies, new non-parametric machine-learning models with higher degrees of freedom, for example, Bayesian Network (e.g., Zhang et al, 2020) and Neural Networks (Chen et al, 2001;Singh et al, 2009;Soltani et al, 2015;Wang et al, 2017) are thriving in the domain of food quality prediction. However, the big data models have not yet been able to replace the traditional parametric quality-decay models governed by pre-specified mathematical functions due to the reasons such as intensive data requirement, and longer computational time.…”