Data-Based Modelling for Prediction

Data-based modelling uses mathematical models fitted from datasets to describe the behaviour of a process or system, in contrast with first-principle models, which provide a fundamental understanding of physicochemical phenomena, such as the ones studied in fluid mechanics, heat transfer, and mass transfer. In data-based modelling, quantitative data analysis methods are used to identify and study a set of variables and determine the relationship or connections between them. This chapter describes simple regression models, non-linear regression models, non-linear machine learning algorithms, and distribution models. We also outline tools for model performance evaluation and validation. Finally, we briefly explore the research methods distinguishing predictive mathematical and causal modelling.