Kinetics of heat-induced changes in dairy products: Developments in data analysis and modelling techniques
M.A.J.S. van Boekel
Abstract
This article describes developments in kinetic modelling of heat-induced chemical changes, protein denaturation, microbial inactivation as a tool for product and process design and shelf life prediction. Computational tools like Bayesian linear and nonlinear regression, multiresponse analysis, and multilevel analysis are reviewed with case studies from literature on dairy products. Bayesian analysis makes uncertainty explicit and pays not only attention to fitting but also to prediction capacity of models. Nonlinear regression obviates the need for data transformation otherwise needed for linear regression. Multiresponse methods allow to build complex reaction networks. Inherent variation of natural products like dairy foods is well characterised by multilevel models, achieved by allowing regression coefficients to vary for different groups. This paper aims to show opportunities for researchers to propose, analyse and test kinetic models, whilst it aims to provide users of kinetic results with background information to critically examine such results.