Evaluation and validation criteria for kinetic models in food science: A critical review

Abstract

Background

Kinetics plays an important role in food science, and statistical tools are indispensable for modeling and identifying suitable models. Various evaluation methods are employed to derive models that describe experimental data with the highest possible accuracy.

Scope and approach

This article critically examines the validation criteria for kinetic models in food science, emphasizing their relevance and application. Key criteria such as residuals, parameter precision, and predictive ability are explored. The article underscores the importance of a thorough understanding of statistical tools to develop accurate kinetic models. Special attention is given to the techniques and criteria used to evaluate these models, with a detailed discussion of their potential and limitations. This study demonstrates the correct use of different methods for evaluating and validating kinetic models, with examples from the literature illustrating the concepts discussed.

Key findings and conclusions

The main outcome is an in-depth analysis of the various methods commonly used in the evaluation of kinetic models, revealing a significant misunderstanding among researchers regarding the true meaning of these tools. Specifically, some well-known tools, as they are currently used, are ineffective, while others, often overlooked, are indispensable. The advantages and disadvantages of these methods are highlighted. The main advantage lies in evaluating uncertainty and predictive capabilities, while the main disadvantage is that researchers need to have a solid understanding of these methods. By addressing gaps in kinetic model evaluation tools, the field of food science could achieve significant advancements.