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.

本文描述了热诱导化学变化、蛋白质变性、微生物灭活的动力学建模的发展，这些建模作为产品和工艺设计以及保质期预测的工具。贝叶斯线性和非线性回归、多响应分析和多级分析等计算工具通过乳制品文献中的案例研究进行了回顾。贝叶斯分析明确了不确定性，不仅关注拟合，还关注模型的预测能力。非线性回归消除了线性回归所需的数据转换的需要。多响应方法允许构建复杂的反应网络。乳制品等天然产品的内在变化可以通过多级模型很好地表征，通过允许不同组的回归系数变化来实现。